Python: [Durabletask] Update feature-durabletask-python branch with main (#3068)

* Python: Add factory pattern to concurrent orchestration builder (#2738) * Add factory pattern to concurrent orchestration builder * Update readme * Address AI comments * Fix unit tests * Fix import * Prevent multiple calls to set participants or factories * Add comments * Mitigate warnings * Fix mypy * Address comments * Address Copilot comments * Fix tests * Python: fix: GroupChat ManagerSelectionResponse JSON Schema for OpenAI Structured Outpu… (#2750) * fix: ManagerSelectionResponse JSON Schema for OpenAI Structured Output Strict Mode * refactor: install pre-commit then commit again * Capture file IDs from code interpreter in streaming responses (#2741) * .NET: [BREAKING] Prevent nulls in AIAgent property (#2719) * prevent nulls in AIAgent property * address feedback * code ql sm04598 (#2723) Co-authored-by: Mark Wallace <127216156+markwallace-microsoft@users.noreply.github.com> * .NET: Add Conversation State Sample (Step05) (#2697) * Initial plan * Add Agent_OpenAI_Step05_Conversation sample for conversation state management Co-authored-by: rogerbarreto <19890735+rogerbarreto@users.noreply.github.com> * Update Program.cs comment to accurately describe the sample Co-authored-by: rogerbarreto <19890735+rogerbarreto@users.noreply.github.com> * Update the code to use the ConversationClient more in line with the samples in OpenAI * Apply suggestions from code review Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> * Changing sample to use ChatClientAgent and conversationId in GetNewThread --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: rogerbarreto <19890735+rogerbarreto@users.noreply.github.com> Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> * Bump AWSSDK.Extensions.Bedrock.MEAI from 4.0.4.7 to 4.0.4.11 (#2777) --- updated-dependencies: - dependency-name: AWSSDK.Extensions.Bedrock.MEAI dependency-version: 4.0.4.11 dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> * Bump Azure.Identity from 1.17.0 to 1.17.1 (#2780) --- updated-dependencies: - dependency-name: Azure.Identity dependency-version: 1.17.1 dependency-type: direct:production update-type: version-update:semver-patch - dependency-name: Azure.Identity dependency-version: 1.17.1 dependency-type: direct:production update-type: version-update:semver-patch - dependency-name: Azure.Identity dependency-version: 1.17.1 dependency-type: direct:production update-type: version-update:semver-patch - dependency-name: Azure.Identity dependency-version: 1.17.1 dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> * Bump Azure.AI.AgentServer.AgentFramework from 1.0.0-beta.4 to 1.0.0-beta.5 (#2778) --- updated-dependencies: - dependency-name: Azure.AI.AgentServer.AgentFramework dependency-version: 1.0.0-beta.5 dependency-type: direct:production update-type: version-update:semver-patch - dependency-name: Azure.AI.AgentServer.AgentFramework dependency-version: 1.0.0-beta.5 dependency-type: direct:production update-type: version-update:semver-patch - dependency-name: Azure.AI.AgentServer.AgentFramework dependency-version: 1.0.0-beta.5 dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> * Python: added more complete parsing for mcp tool arguments (#2756) * added more complete parsing for mcp tool arguments * fixed mypy * added nonlocal model counter, and some fixes * fixes in naming logic * extracted json parsing function, added parametrized test and checked coverage * Python: Updated package versions (#2784) * Updated package versions * Small fix * Bump actions/checkout from 5 to 6 (#2404) Bumps [actions/checkout](https://github.com/actions/checkout) from 5 to 6. - [Release notes](https://github.com/actions/checkout/releases) - [Changelog](https://github.com/actions/checkout/blob/main/CHANGELOG.md) - [Commits](https://github.com/actions/checkout/compare/v5...v6) --- updated-dependencies: - dependency-name: actions/checkout dependency-version: '6' dependency-type: direct:production update-type: version-update:semver-major ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Chris <66376200+crickman@users.noreply.github.com> * .NET: adds support for labels in edges, fixes rendering of labels in dot a… (#1507) * adds support for labels in edges, fixes rendering of labels in dot and mermaid, adds rendering of labels in edges * Update dotnet/src/Microsoft.Agents.AI.Workflows/Visualization/WorkflowVisualizer.cs Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> * escaping edge labels, adding tests for labels containing strange characters that would break the diagram and enabling the previous signature so the API has backwards compatibility. * Unify label in EdgeData * Edge API adjustments, removed useless "sanitizer" * fixed test --------- Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> Co-authored-by: Jacob Alber <jaalber@microsoft.com> Co-authored-by: Chris <66376200+crickman@users.noreply.github.com> * Python: Added custom args and thread object to ai_function kwargs (#2769) * Added an example of using kwargs in ai_function * Added thread object to ai_function kwargs * Updated docs * Small fix * Added thread parameter filtering * Fix WorkflowAgent to include thread convo history. Enable checkpointing. (#2774) * Update OpenAIResponses.yaml to match AgentSchema (#2598) 1. Update `connection` child types -- `kind: ApiKey` to `kind: key` otherwise schema will fail: https://microsoft.github.io/AgentSchema/reference/apikeyconnection/ 2. Update `outputSchema`'s `PropertySchema` to be `kind` instead of `type` otherwise schema will fail: https://microsoft.github.io/AgentSchema/reference/propertyschema/ * Python: Remove warnings from workflow builder on not using factories (#2808) * Revert concurrent * Fix comments * Python: Filter framework kwargs from MCP tool invocations (#2870) * Filter framework kwargs from MCP tool invocations * Fixes * Python: Fix WorkflowAgent to emit yield_output as agent response (#2866) * Fix WorkflowAgent to emit yield_output as agent response * use raw_representation * Raw representation handling * Python: Use agent description in HandoffBuilder auto-generated tools (#2713) (#2714) ## Summary Enhanced `HandoffBuilder._apply_auto_tools` to use the target agent's description when creating handoff tools, providing more informative tool descriptions for LLMs. ## Changes - Modified `_apply_auto_tools` to extract `description` from `AgentExecutor._agent` when available - Updated iteration to use `.items()` for more efficient dict traversal - Handoff tools now use agent descriptions instead of generic placeholders ## Example Before: "Handoff to the refund_agent agent." After: "You handle refund requests. Ask for order details and process refunds." ## Testing - All handoff tests pass (20/20) - No breaking changes to existing API Fixes #2713 Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com> * Python: [BREAKING] Observability updates (#2782) * fixes Python: Add env_file_path parameter to setup_observability() similar to AzureOpenAIChatClient Fixes #2186 * WIP on updates using configure_azure_monitor * improved setup and clarity * fixed root .env.example * revert changes * updated files * updated sample * updated zero code * test fixes and fixed links * fix devui * removed planning docs * added enable method and updated readme and samples * clarified docstring * add return annotation * updated naming * update capatilized version * updated readme and some fixes * updated decorator name inline with the rest * feedback from comments addressed * Python: Fix middleware terminate flag to exit function calling loop immediately (#2868) * Fix middleware terminate flag to exit function calling loop immediately * Eliminating duck typing * Improve function exec result handling * Fix race condition * Fix mypy issues * Python: Fix context duplication in handoff workflows when restoring from checkpoint (#2867) * Fix context duplication in handoff workflows when restoring from checkpoint * Address Copilot PR review * .NET: Update to latest Azure.AI.*, OpenAI, and M.E.AI* (#2850) * Update to latest Azure.AI.*, OpenAI, and M.E.AI* Absorb breaking changes in Responses surface area * Update dotnet/samples/AgentWebChat/AgentWebChat.AgentHost/Utilities/ChatClientExtensions.cs * Update dotnet/samples/AgentWebChat/AgentWebChat.AgentHost/Utilities/ChatClientExtensions.cs * Update dotnet/samples/AgentWebChat/AgentWebChat.AgentHost/Utilities/ChatClientExtensions.cs * Update dotnet/samples/GettingStarted/AgentWithOpenAI/Agent_OpenAI_Step04_CreateFromOpenAIResponseClient/Program.cs Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> * Using patch to remove the model is necessary, updated the response client to actually use the the ForAgent --------- Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> Co-authored-by: Roger Barreto <19890735+rogerbarreto@users.noreply.github.com> * Bump actions/download-artifact from 6 to 7 (#2862) Bumps [actions/download-artifact](https://github.com/actions/download-artifact) from 6 to 7. - [Release notes](https://github.com/actions/download-artifact/releases) - [Commits](https://github.com/actions/download-artifact/compare/v6...v7) --- updated-dependencies: - dependency-name: actions/download-artifact dependency-version: '7' dependency-type: direct:production update-type: version-update:semver-major ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> * Bump actions/cache from 4 to 5 (#2861) Bumps [actions/cache](https://github.com/actions/cache) from 4 to 5. - [Release notes](https://github.com/actions/cache/releases) - [Changelog](https://github.com/actions/cache/blob/main/RELEASES.md) - [Commits](https://github.com/actions/cache/compare/v4...v5) --- updated-dependencies: - dependency-name: actions/cache dependency-version: '5' dependency-type: direct:production update-type: version-update:semver-major ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> * Bump actions/upload-artifact from 5 to 6 (#2860) Bumps [actions/upload-artifact](https://github.com/actions/upload-artifact) from 5 to 6. - [Release notes](https://github.com/actions/upload-artifact/releases) - [Commits](https://github.com/actions/upload-artifact/compare/v5...v6) --- updated-dependencies: - dependency-name: actions/upload-artifact dependency-version: '6' dependency-type: direct:production update-type: version-update:semver-major ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> * Python : Ollama Connector for Agent Framework (#1104) * Initial Commit for Olama Connector * Added Olama Sample * Add Sample & Fixed Open Telemetry * Fixed Spelling from Olama to Ollama * remove"opentelemetry-semantic-conventions-ai ~=0.4.13" since its handled in a different pr * Added Tool Calling * Finalizing test cases * Adjust samples to be more reliable * Update python/packages/ollama/agent_framework_ollama/_chat_client.py Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> * Update python/packages/ollama/pyproject.toml Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> * Update python/packages/ollama/tests/test_ollama_chat_client.py Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> * Update python/packages/ollama/agent_framework_ollama/_chat_client.py Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> * Improved Docstrings & Sample * Update python/packages/ollama/agent_framework_ollama/_chat_client.py Co-authored-by: Eduard van Valkenburg <eavanvalkenburg@users.noreply.github.com> * Integrate PR Feedback - Divided Streaming and Non-Streaming into independent Methods - Catch Ollama Validation Error - Add OTEL Provider Name - Checked Ollama Messages - Add Usage Statistics * Revert setting, so it can be none * Validate Message formatting between AF and Ollama * Catch Ollama Error and raise a ServiceResponse Error * Fix mypy error * remove .vscode comma * Add Reasoning support & adjust to new structure * Add Ollama Multimodality and Reasoning * Add test cases for reasoning * Add Tests for Error Handling in Ollama Client * Update python/samples/getting_started/multimodal_input/ollama_chat_multimodal.py Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> * Integrated Copilot Feedback * Implement first PR Feedback * Adjust Readme files for examples * Adjust argument passing via additional chat options * Implemented PR Feedback * Removing Ollama Package from Core and moving samples * Fix Link & Adding Samples to Main Sample Readme * Fixing Links in Readme * Moved Multimodal and Chat Example * Fixed Link in ChatClient to Ollama * Fix AgentFramework Links in Ollama Project * Fix observability breaking change --------- Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> Co-authored-by: Eduard van Valkenburg <eavanvalkenburg@users.noreply.github.com> * Skip failing IT (#2904) * .NET: Cosmos DB UT Fast Skip (For Non-Configured Local envs) (#2906) * Cosmos DB UT Fast Skip (Non-Configured Local envs) + Long running UT skip in pipeline when no CosmosDB changes happened * Force a CosmosDB source code change to trigger the pipeline * Address possible string boolean mismatch * Add debug * Enabling emulator always when running IT * .NET: Add TTLs to durable agent sessions (#2679) * .NET: Add TTLs to durable agent sessions * Remove unnecessary async * PR feedback: clarify UTC * PR feedback: limit minimum signal delay to <= 5 minutes * PR feedback: Fix TTL disablement * Linter: use auto-property * Fix build break from OpenAI SDK change * Updated CHANGELOG.md * PR feedback * Reduce default TTL to 14 days to work around DTS bug * Python: Update Mem0Provider to use v2 search API `filters` parameter (#2766) * short fix to move id parameters to filters object * added tests * small fix * mem0 dependency update * Updated package versions (#2913) * .NET: Switch to new "Run" method name. (#2843) * Switch to new "RunAgent" method name. * Try to disable false positive naming warning. * Add comment about disabled warnings. * Rename `RunAgent` to just `Run`. * Update CHANGELOG. * Python: Switch to new "run" method name. (#2890) * Switch to `run` method. * Add support for deprecated `run_agent`. * Fix entity method name. * Fix method name and improve tests. * Update comment. * Update Python CHANGELOG. * [BREAKING] Python: Add factory pattern to handoff orchestration builder (#2844) * WIP: Factory pattern to handoff * Add factory pattern to concurrent orchestration builder; Next: tests and sample verification * Add tests and improve comments * Fix mypy * Simplify handoff_simple.py * Simplify handoff_autonoumous.py and bug fix * Update readme * Address Copilot comments * Python: Flow custom kwargs to agents via Workflow SharedState (#2894) * Flow custom kwargs to agents via SharedState * Address Copilot feedback * Improve sample typing * Fix test * Fix Pydantic error when using Literal type for tool params (#2893) * Updated Ollama package version (#2920) * Python: Azure AI Agent with Bing Grounding Citations Sample (#2892) * bing grounding sample with citations * small fix * fix * .NET: Make DelegatingAIAgent abstract (#2797) * Initial plan * Make DelegatingAIAgent abstract Co-authored-by: SergeyMenshykh <68852919+SergeyMenshykh@users.noreply.github.com> --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: SergeyMenshykh <68852919+SergeyMenshykh@users.noreply.github.com> * Added additional arguments for Azure AI agent (#2922) * Python: Correction of MCP image type conversion in _mcp.py (#2901) * Correction of MCP image type conversion in _mcp.py * Added a new overload to the init function of the DataContent() type of the Agent Framework, edited the test case to correctly test the usage of the data and uri fields while using DataContent() * Fixed tests related to the changes of the DataContent type, added testing for both string and byte representations * Pass kwargs into subworkflows (#2923) * Python: Move ollama samples to samples getting started dir (#2921) * Move ollama samples to samples getting started dir * Address feedback * Python: fix: correct BadRequestError when using Pydantic model in response_fo… (#1843) * fix: correct BadRequestError when using Pydantic model in response_format * Fix lint --------- Co-authored-by: Evan Mattson <evan.mattson@microsoft.com> * .NET: [Breaking] Delete display name property (#2758) * delete the AIAgent.DisplayName property * use agent name as a first value for activity display name * Update dotnet/src/Microsoft.Agents.AI.Workflows/Specialized/HandoffAgentExecutor.cs Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> --------- Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> * Python: cleanup and refactoring of chat clients (#2937) * refactoring and unifying naming schemes of internal methods of chat clients * set tool_choice to auto * fix for mypy * added note on naming and fix #2951 * fix responses * fixes in azure ai agents client * Python: Workflow add option to visualize internal executors (#2917) * Workflow add option to visualize internal executors * Address Copilot comments * Python: Fixes Run ID and Thread ID casing to align with AG-UI Typescript SDK (#2948) * added camelCase input to run id and thread id aligning with @ag-ui/core * fixed per copilot suggestions * Python: Add workflow cancellation sample (#2732) * Add workflow cancellation sample Add sample demonstrating how to cancel a running workflow using asyncio tasks. Shows both cancellation mid-execution and normal completion paths. Useful for implementing timeouts, graceful shutdown, or A2A executors. * update docstring * .NET: Update Anthropic package to version 12.0.0 (#2914) * Initial plan * Update Anthropic package to version 12.0.0 Co-authored-by: stephentoub <2642209+stephentoub@users.noreply.github.com> --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: stephentoub <2642209+stephentoub@users.noreply.github.com> * Python: Add Azure Managed Redis Support with Credential Provider (#2887) * azure redis support * small fixes * azure managed redis sample * fixes * Bump CommunityToolkit.Aspire.OllamaSharp from 13.0.0-beta.440 to 13.0.0 (#2856) --- updated-dependencies: - dependency-name: CommunityToolkit.Aspire.OllamaSharp dependency-version: 13.0.0 dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> * Bump AWSSDK.Extensions.Bedrock.MEAI from 4.0.4.11 to 4.0.5 (#2853) --- updated-dependencies: - dependency-name: AWSSDK.Extensions.Bedrock.MEAI dependency-version: 4.0.5 dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Mark Wallace <127216156+markwallace-microsoft@users.noreply.github.com> * Bump Azure.AI.AgentServer.AgentFramework from 1.0.0-beta.4 to 1.0.0-beta.5 (#2854) --- updated-dependencies: - dependency-name: Azure.AI.AgentServer.AgentFramework dependency-version: 1.0.0-beta.5 dependency-type: direct:production update-type: version-update:semver-patch - dependency-name: Azure.AI.AgentServer.AgentFramework dependency-version: 1.0.0-beta.5 dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Chris <66376200+crickman@users.noreply.github.com> * Python: Fix WorkflowAgent event handling and kwargs forwarding (#2946) * Fix kwargs propagation through workflow.as_agent() * Fix WorkflowAgent to respect AgentExecutor output_response setting * .NET: Use GrpcEntityRunner instead of TaskEntityDispatcher (#2759) * Use GrpcEntityRunner instead of TaskEntityDispatcher * Pin to Durable worker 1.11.0 * Set the invocation result * Update all Durable packages * Update changelog, rename dispatcher to encondedEntityRequest * Python: Bump Py version to 1.0.0b251218 for a release. Update CHANGELOG (#2968) * Bump Py version to 1.0.0b251218 for a release. Update CHANGELOG * update lock * Fix formatting * Fix ChatKit typing * Python: Introducing Foundry Local Chat Clients (#2915) * redo foundry local chat client * fix mypy and spelling * better docstring, updated sample * fixed tests and added tests * small sample update * Updated package versions (#2978) * Python: Added GitHub MCP sample with PAT (#2967) * added github mcp sample with PAT * addressed copilot fixes * env fix * Python: Preserve reasoning blocks with OpenRouter (#2950) * Preserve reasoning blocks with OpenRouter * Put encrypted reasoning in TextReasoningContent * Remove unneccessary change * Fix docs * Support streaming * Fix handling None in TextReasoningContent.text * Python: Added response.created and response.in_progress event process to OpenAIBaseResponseClient (#2975) * added response.created and response.in_progress to include response.id * better doc string * added tests for the new streaming event types * Python: Introducing support for Bedrock-hosted models (Anthropic, Cohere, etc.) (#2610) * Pushing the bedrock related changes to the new branch after addressing the review comments * 2524 Addressed the second round review comments * 2524 Addressed few more minor comments on the PR * resolving the merge conflict * 2524 resolved the uv.lock conflicts * 2524 addressed more comments * 2524 removed the print statement to fix the checks failure * 2524 resolved the CI failure issues * 2524 fixing the CI breaks * 2524 Addressed the review comment * 2524 resolved conflict --------- Co-authored-by: Sunil Dutta <sunil.dutta@penske.com> Co-authored-by: budgetboardingai <apurva.sharma31@gmail.com> * .NET: [Durable Agents] Reliable streaming sample (#2942) * .NET: [Durable Agents] Reliable streaming sample * Add automated validation for new sample * Address Copilot PR feedback * Fix typo in README.md about agent definitions (#2634) * Fix typo in README.md about agent definitions * Update agent-samples/README.md Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> --------- Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com> Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> * Python: latency improvements (#3014) * latency improvements * fixed mypy, added coding standards and instructions * slight logic improvement * Python: Updated package versions (#3024) * Updated package versions * Updated changelog * Python: add powerfx safe mode (#3028) * add powerfx safe mode * improved docstring and aligned env_file loading * ensured test uses reset * .NET: [Breaking] Introduce RunCoreAsync/RunCoreStreamingAsync delegation pattern in AIAgent (#2749) * Initial plan * Refactor AIAgent: Make RunAsync and RunStreamingAsync non-abstract, add RunCoreAsync and RunCoreStreamingAsync Co-authored-by: SergeyMenshykh <68852919+SergeyMenshykh@users.noreply.github.com> * Fix infinite recursion in test implementations Co-authored-by: SergeyMenshykh <68852919+SergeyMenshykh@users.noreply.github.com> * Make RunAsync and RunStreamingAsync non-virtual as requested Co-authored-by: SergeyMenshykh <68852919+SergeyMenshykh@users.noreply.github.com> * Fix DelegatingAIAgent subclasses to use RunCoreAsync/RunCoreStreamingAsync Co-authored-by: SergeyMenshykh <68852919+SergeyMenshykh@users.noreply.github.com> * Fix XML documentation references in AnonymousDelegatingAIAgent Co-authored-by: SergeyMenshykh <68852919+SergeyMenshykh@users.noreply.github.com> * Restore <see cref> tags with proper qualified signatures in AnonymousDelegatingAIAgent Co-authored-by: SergeyMenshykh <68852919+SergeyMenshykh@users.noreply.github.com> * Rollback unnecessary XML documentation changes in AnonymousDelegatingAIAgent Co-authored-by: SergeyMenshykh <68852919+SergeyMenshykh@users.noreply.github.com> * Remove pragma and update crefs to RunCoreAsync/RunCoreStreamingAsync Co-authored-by: SergeyMenshykh <68852919+SergeyMenshykh@users.noreply.github.com> * Fix EntityAgentWrapper to call base.RunCoreAsync/RunCoreStreamingAsync Co-authored-by: SergeyMenshykh <68852919+SergeyMenshykh@users.noreply.github.com> * fix compilation issues * fix compilatio issue * fix tests * fix unit tests * fix unit test --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: SergeyMenshykh <68852919+SergeyMenshykh@users.noreply.github.com> Co-authored-by: SergeyMenshykh <sergemenshikh@gmail.com> Co-authored-by: Chris <66376200+crickman@users.noreply.github.com> * Remove from feature branch * Remove ollama changes --------- Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: Tao Chen <taochen@microsoft.com> Co-authored-by: Kurt <65111699+q33566@users.noreply.github.com> Co-authored-by: Evan Mattson <35585003+moonbox3@users.noreply.github.com> Co-authored-by: SergeyMenshykh <68852919+SergeyMenshykh@users.noreply.github.com> Co-authored-by: Korolev Dmitry <deagle.gross@gmail.com> Co-authored-by: Mark Wallace <127216156+markwallace-microsoft@users.noreply.github.com> Co-authored-by: Copilot <198982749+Copilot@users.noreply.github.com> Co-authored-by: rogerbarreto <19890735+rogerbarreto@users.noreply.github.com> Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Eduard van Valkenburg <eavanvalkenburg@users.noreply.github.com> Co-authored-by: Dmytro Struk <13853051+dmytrostruk@users.noreply.github.com> Co-authored-by: Chris <66376200+crickman@users.noreply.github.com> Co-authored-by: Jose Luis Latorre Millas <joslat@gmail.com> Co-authored-by: Jacob Alber <jaalber@microsoft.com> Co-authored-by: Richard Ortega <richardjortega@gmail.com> Co-authored-by: 刘邦学AI <lbbniu@gmail.com> Co-authored-by: Stephen Toub <stoub@microsoft.com> Co-authored-by: Nico Möller <nkm-moeller@mail.de> Co-authored-by: Chris Gillum <cgillum@microsoft.com> Co-authored-by: Giles Odigwe <79032838+giles17@users.noreply.github.com> Co-authored-by: Phillip Hoff <phillip.hoff@gmail.com> Co-authored-by: Ege Ozan Özyedek <36128615+egeozanozyedek@users.noreply.github.com> Co-authored-by: samueljohnsiby <66901393+samueljohnsiby@users.noreply.github.com> Co-authored-by: Evan Mattson <evan.mattson@microsoft.com> Co-authored-by: Hao Luo <338265+howlowck@users.noreply.github.com> Co-authored-by: Victor Dibia <chuvidi2003@gmail.com> Co-authored-by: stephentoub <2642209+stephentoub@users.noreply.github.com> Co-authored-by: Jacob Viau <javia@microsoft.com> Co-authored-by: SuperKenVery <39673849+SuperKenVery@users.noreply.github.com> Co-authored-by: Sunil Dutta <dutta.2003@gmail.com> Co-authored-by: Sunil Dutta <sunil.dutta@penske.com> Co-authored-by: budgetboardingai <apurva.sharma31@gmail.com> Co-authored-by: Syrine Chelly <62653967+SyChell@users.noreply.github.com> Co-authored-by: SergeyMenshykh <sergemenshikh@gmail.com>
2026-06-16 21:04:09 +08:00 · 2026-01-02 13:13:37 -08:00
parent a02527f00a
commit a5b36dc379
182 changed files with 7454 additions and 2462 deletions
@@ -501,7 +501,7 @@ class BaseChatClient(SerializationMixin, ABC):
        stop: str | Sequence[str] | None = None,
        store: bool | None = None,
        temperature: float | None = None,
-        tool_choice: ToolMode | Literal["auto", "required", "none"] | dict[str, Any] | None = None,
+        tool_choice: ToolMode | Literal["auto", "required", "none"] | dict[str, Any] | None = "auto",
        tools: ToolProtocol
        | Callable[..., Any]
        | MutableMapping[str, Any]
@@ -535,6 +535,7 @@ class BaseChatClient(SerializationMixin, ABC):
            store: Whether to store the response.
            temperature: The sampling temperature to use.
            tool_choice: The tool choice for the request.
+                Default is `auto`.
            tools: The tools to use for the request.
            top_p: The nucleus sampling probability to use.
            user: The user to associate with the request.
@@ -595,7 +596,7 @@ class BaseChatClient(SerializationMixin, ABC):
        stop: str | Sequence[str] | None = None,
        store: bool | None = None,
        temperature: float | None = None,
-        tool_choice: ToolMode | Literal["auto", "required", "none"] | dict[str, Any] | None = None,
+        tool_choice: ToolMode | Literal["auto", "required", "none"] | dict[str, Any] | None = "auto",
        tools: ToolProtocol
        | Callable[..., Any]
        | MutableMapping[str, Any]
@@ -629,6 +630,7 @@ class BaseChatClient(SerializationMixin, ABC):
            store: Whether to store the response.
            temperature: The sampling temperature to use.
            tool_choice: The tool choice for the request.
+                Default is `auto`.
            tools: The tools to use for the request.
            top_p: The nucleus sampling probability to use.
            user: The user to associate with the request.
@@ -63,21 +63,21 @@ __all__ = [
 ]


-def _mcp_prompt_message_to_chat_message(
+def _parse_message_from_mcp(
    mcp_type: types.PromptMessage | types.SamplingMessage,
 ) -> ChatMessage:
-    """Convert a MCP container type to a Agent Framework type."""
+    """Parse an MCP container type into an Agent Framework type."""
    return ChatMessage(
        role=Role(value=mcp_type.role),
-        contents=_mcp_type_to_ai_content(mcp_type.content),
+        contents=_parse_content_from_mcp(mcp_type.content),
        raw_representation=mcp_type,
    )


-def _mcp_call_tool_result_to_ai_contents(
+def _parse_contents_from_mcp_tool_result(
    mcp_type: types.CallToolResult,
 ) -> list[Contents]:
-    """Convert a MCP container type to a Agent Framework type.
+    """Parse an MCP CallToolResult into Agent Framework content types.

    This function extracts the complete _meta field from CallToolResult objects
    and merges all metadata into the additional_properties field of converted
@@ -111,7 +111,7 @@ def _mcp_call_tool_result_to_ai_contents(
    # Convert each content item and merge metadata
    result_contents = []
    for item in mcp_type.content:
-        contents = _mcp_type_to_ai_content(item)
+        contents = _parse_content_from_mcp(item)

        if merged_meta_props:
            for content in contents:
@@ -124,7 +124,7 @@ def _mcp_call_tool_result_to_ai_contents(
    return result_contents


-def _mcp_type_to_ai_content(
+def _parse_content_from_mcp(
    mcp_type: types.ImageContent
    | types.TextContent
    | types.AudioContent
@@ -142,7 +142,7 @@ def _mcp_type_to_ai_content(
        | types.ToolResultContent
    ],
 ) -> list[Contents]:
-    """Convert a MCP type to a Agent Framework type."""
+    """Parse an MCP type into an Agent Framework type."""
    mcp_types = mcp_type if isinstance(mcp_type, Sequence) else [mcp_type]
    return_types: list[Contents] = []
    for mcp_type in mcp_types:
@@ -178,7 +178,7 @@ def _mcp_type_to_ai_content(
                return_types.append(
                    FunctionResultContent(
                        call_id=mcp_type.toolUseId,
-                        result=_mcp_type_to_ai_content(mcp_type.content)
+                        result=_parse_content_from_mcp(mcp_type.content)
                        if mcp_type.content
                        else mcp_type.structuredContent,
                        exception=Exception() if mcp_type.isError else None,
@@ -211,10 +211,10 @@ def _mcp_type_to_ai_content(
    return return_types


-def _ai_content_to_mcp_types(
+def _prepare_content_for_mcp(
    content: Contents,
 ) -> types.TextContent | types.ImageContent | types.AudioContent | types.EmbeddedResource | types.ResourceLink | None:
-    """Convert a BaseContent type to a MCP type."""
+    """Prepare an Agent Framework content type for MCP."""
    match content:
        case TextContent():
            return types.TextContent(type="text", text=content.text)
@@ -253,15 +253,15 @@ def _ai_content_to_mcp_types(
            return None


-def _chat_message_to_mcp_types(
+def _prepare_message_for_mcp(
    content: ChatMessage,
 ) -> list[types.TextContent | types.ImageContent | types.AudioContent | types.EmbeddedResource | types.ResourceLink]:
-    """Convert a ChatMessage to a list of MCP types."""
+    """Prepare a ChatMessage for MCP format."""
    messages: list[
        types.TextContent | types.ImageContent | types.AudioContent | types.EmbeddedResource | types.ResourceLink
    ] = []
    for item in content.contents:
-        mcp_content = _ai_content_to_mcp_types(item)
+        mcp_content = _prepare_content_for_mcp(item)
        if mcp_content:
            messages.append(mcp_content)
    return messages
@@ -469,7 +469,7 @@ class MCPTool:
        logger.debug("Sampling callback called with params: %s", params)
        messages: list[ChatMessage] = []
        for msg in params.messages:
-            messages.append(_mcp_prompt_message_to_chat_message(msg))
+            messages.append(_parse_message_from_mcp(msg))
        try:
            response = await self.chat_client.get_response(
                messages,
@@ -487,7 +487,7 @@ class MCPTool:
                code=types.INTERNAL_ERROR,
                message="Failed to get chat message content.",
            )
-        mcp_contents = _chat_message_to_mcp_types(response.messages[0])
+        mcp_contents = _prepare_message_for_mcp(response.messages[0])
        # grab the first content that is of type TextContent or ImageContent
        mcp_content = next(
            (content for content in mcp_contents if isinstance(content, (types.TextContent, types.ImageContent))),
@@ -692,7 +692,7 @@ class MCPTool:
            k: v for k, v in kwargs.items() if k not in {"chat_options", "tools", "tool_choice", "thread"}
        }
        try:
-            return _mcp_call_tool_result_to_ai_contents(
+            return _parse_contents_from_mcp_tool_result(
                await self.session.call_tool(tool_name, arguments=filtered_kwargs)
            )
        except McpError as mcp_exc:
@@ -724,7 +724,7 @@ class MCPTool:
            )
        try:
            prompt_result = await self.session.get_prompt(prompt_name, arguments=kwargs)
-            return [_mcp_prompt_message_to_chat_message(message) for message in prompt_result.messages]
+            return [_parse_message_from_mcp(message) for message in prompt_result.messages]
        except McpError as mcp_exc:
            raise ToolExecutionException(mcp_exc.error.message, inner_exception=mcp_exc) from mcp_exc
        except Exception as ex:
@@ -573,7 +573,7 @@ class AIFunction(BaseTool, Generic[ArgsT, ReturnT]):
    """

    INJECTABLE: ClassVar[set[str]] = {"func"}
-    DEFAULT_EXCLUDE: ClassVar[set[str]] = {"input_model", "_invocation_duration_histogram"}
+    DEFAULT_EXCLUDE: ClassVar[set[str]] = {"input_model", "_invocation_duration_histogram", "_cached_parameters"}

    def __init__(
        self,
@@ -615,6 +615,7 @@ class AIFunction(BaseTool, Generic[ArgsT, ReturnT]):
        self.func = func
        self._instance = None  # Store the instance for bound methods
        self.input_model = self._resolve_input_model(input_model)
+        self._cached_parameters: dict[str, Any] | None = None  # Cache for model_json_schema()
        self.approval_mode = approval_mode or "never_require"
        if max_invocations is not None and max_invocations < 1:
            raise ValueError("max_invocations must be at least 1 or None.")
@@ -802,8 +803,11 @@ class AIFunction(BaseTool, Generic[ArgsT, ReturnT]):

        Returns:
            A dictionary containing the JSON schema for the function's parameters.
+            The result is cached after the first call for performance.
        """
-        return self.input_model.model_json_schema()
+        if self._cached_parameters is None:
+            self._cached_parameters = self.input_model.model_json_schema()
+        return self._cached_parameters

    def to_json_schema_spec(self) -> dict[str, Any]:
        """Convert a AIFunction to the JSON Schema function specification format.
@@ -825,7 +829,7 @@ class AIFunction(BaseTool, Generic[ArgsT, ReturnT]):
        as_dict = super().to_dict(exclude=exclude, exclude_none=exclude_none)
        if (exclude and "input_model" in exclude) or not self.input_model:
            return as_dict
-        as_dict["input_model"] = self.input_model.model_json_schema()
+        as_dict["input_model"] = self.parameters()  # Use cached parameters()
        return as_dict


@@ -1779,11 +1783,6 @@ def _handle_function_calls_response(
            response: "ChatResponse | None" = None
            fcc_messages: "list[ChatMessage]" = []

-            # If tools are provided but tool_choice is not set, default to "auto" for function invocation
-            tools = _extract_tools(kwargs)
-            if tools and kwargs.get("tool_choice") is None:
-                kwargs["tool_choice"] = "auto"
-
            for attempt_idx in range(config.max_iterations if config.enabled else 0):
                fcc_todo = _collect_approval_responses(prepped_messages)
                if fcc_todo:
@@ -101,7 +101,7 @@ def _parse_content(content_data: MutableMapping[str, Any]) -> "Contents":
    Raises:
        ContentError if parsing fails
    """
-    content_type = str(content_data.get("type"))
+    content_type: str | None = content_data.get("type", None)
    match content_type:
        case "text":
            return TextContent.from_dict(content_data)
@@ -127,6 +127,8 @@ def _parse_content(content_data: MutableMapping[str, Any]) -> "Contents":
            return FunctionApprovalResponseContent.from_dict(content_data)
        case "text_reasoning":
            return TextReasoningContent.from_dict(content_data)
+        case None:
+            raise ContentError("Content type is missing")
        case _:
            raise ContentError(f"Unknown content type '{content_type}'")

@@ -789,8 +791,9 @@ class TextReasoningContent(BaseContent):

    def __init__(
        self,
-        text: str,
+        text: str | None,
        *,
+        protected_data: str | None = None,
        additional_properties: dict[str, Any] | None = None,
        raw_representation: Any | None = None,
        annotations: Sequence[Annotations | MutableMapping[str, Any]] | None = None,
@@ -802,6 +805,16 @@ class TextReasoningContent(BaseContent):
            text: The text content represented by this instance.

        Keyword Args:
+            protected_data: This property is used to store data from a provider that should be roundtripped back to the
+                provider but that is not intended for human consumption. It is often encrypted or otherwise redacted
+                information that is only intended to be sent back to the provider and not displayed to the user. It's
+                possible for a TextReasoningContent to contain only `protected_data` and have an empty `text` property.
+                This data also may be associated with the corresponding `text`, acting as a validation signature for it.
+
+                Note that whereas `text` can be provider agnostic, `protected_data` is provider-specific, and is likely
+                to only be understood by the provider that created it. The data is often represented as a more complex
+                object, so it should be serialized to a string before storing so that the whole object is easily
+                serializable without loss.
            additional_properties: Optional additional properties associated with the content.
            raw_representation: Optional raw representation of the content.
            annotations: Optional annotations associated with the content.
@@ -814,6 +827,7 @@ class TextReasoningContent(BaseContent):
            **kwargs,
        )
        self.text = text
+        self.protected_data = protected_data
        self.type: Literal["text_reasoning"] = "text_reasoning"

    def __add__(self, other: "TextReasoningContent") -> "TextReasoningContent":
@@ -846,13 +860,18 @@ class TextReasoningContent(BaseContent):
        else:
            annotations = self.annotations + other.annotations

+        # Replace protected data.
+        # Discussion: https://github.com/microsoft/agent-framework/pull/2950#discussion_r2634345613
+        protected_data = other.protected_data or self.protected_data
+
        # Create new instance using from_dict for proper deserialization
        result_dict = {
-            "text": self.text + other.text,
+            "text": (self.text or "") + (other.text or "") if self.text is not None or other.text is not None else None,
            "type": "text_reasoning",
            "annotations": [ann.to_dict(exclude_none=False) for ann in annotations] if annotations else None,
            "additional_properties": {**(self.additional_properties or {}), **(other.additional_properties or {})},
            "raw_representation": raw_representation,
+            "protected_data": protected_data,
        }
        return TextReasoningContent.from_dict(result_dict)

@@ -869,7 +888,9 @@ class TextReasoningContent(BaseContent):
            raise TypeError("Incompatible type")

        # Concatenate text
-        self.text += other.text
+        if self.text is not None or other.text is not None:
+            self.text = (self.text or "") + (other.text or "")
+        # if both are None, should keep as None

        # Merge additional properties (self takes precedence)
        if self.additional_properties is None:
@@ -888,6 +909,11 @@ class TextReasoningContent(BaseContent):
                self.raw_representation if isinstance(self.raw_representation, list) else [self.raw_representation]
            ) + (other.raw_representation if isinstance(other.raw_representation, list) else [other.raw_representation])

+        # Replace protected data.
+        # Discussion: https://github.com/microsoft/agent-framework/pull/2950#discussion_r2634345613
+        if other.protected_data is not None:
+            self.protected_data = other.protected_data
+
        # Merge annotations
        if other.annotations:
            if self.annotations is None:
@@ -2224,27 +2250,30 @@ def _process_update(
    if update.message_id:
        message.message_id = update.message_id
    for content in update.contents:
-        if (
-            isinstance(content, FunctionCallContent)
-            and len(message.contents) > 0
-            and isinstance(message.contents[-1], FunctionCallContent)
-        ):
+        # Fast path: get type attribute (most content will have it)
+        content_type = getattr(content, "type", None)
+        # Slow path: only check for dict if type is None
+        if content_type is None and isinstance(content, (dict, MutableMapping)):
            try:
-                message.contents[-1] += content
-            except AdditionItemMismatch:
-                message.contents.append(content)
-        elif isinstance(content, UsageContent):
-            if response.usage_details is None:
-                response.usage_details = UsageDetails()
-            response.usage_details += content.details
-        elif isinstance(content, (dict, MutableMapping)):
-            try:
-                cont = _parse_content(content)
-                message.contents.append(cont)
+                content = _parse_content(content)
+                content_type = content.type
            except ContentError as exc:
                logger.warning(f"Skipping unknown content type or invalid content: {exc}")
-        else:
-            message.contents.append(content)
+                continue
+        match content_type:
+            # mypy doesn't narrow type based on match/case, but we know these are FunctionCallContents
+            case "function_call" if message.contents and message.contents[-1].type == "function_call":
+                try:
+                    message.contents[-1] += content  # type: ignore[operator]
+                except AdditionItemMismatch:
+                    message.contents.append(content)
+            case "usage":
+                if response.usage_details is None:
+                    response.usage_details = UsageDetails()
+                # mypy doesn't narrow type based on match/case, but we know this is UsageContent
+                response.usage_details += content.details  # type: ignore[union-attr, arg-type]
+            case _:
+                message.contents.append(content)
    # Incorporate the update's properties into the response.
    if update.response_id:
        response.response_id = update.response_id
@@ -26,6 +26,7 @@ from agent_framework import (
 )

 from ..exceptions import AgentExecutionException
+from ._agent_executor import AgentExecutor
 from ._checkpoint import CheckpointStorage
 from ._events import (
    AgentRunUpdateEvent,
@@ -141,7 +142,8 @@ class WorkflowAgent(BaseAgent):
            checkpoint_storage: Runtime checkpoint storage. When provided with checkpoint_id,
                used to load and restore the checkpoint. When provided without checkpoint_id,
                enables checkpointing for this run.
-            **kwargs: Additional keyword arguments.
+            **kwargs: Additional keyword arguments passed through to underlying workflow
+                and ai_function tools.

        Returns:
            The final workflow response as an AgentRunResponse.
@@ -153,7 +155,7 @@ class WorkflowAgent(BaseAgent):
        response_id = str(uuid.uuid4())

        async for update in self._run_stream_impl(
-            input_messages, response_id, thread, checkpoint_id, checkpoint_storage
+            input_messages, response_id, thread, checkpoint_id, checkpoint_storage, **kwargs
        ):
            response_updates.append(update)

@@ -187,7 +189,8 @@ class WorkflowAgent(BaseAgent):
            checkpoint_storage: Runtime checkpoint storage. When provided with checkpoint_id,
                used to load and restore the checkpoint. When provided without checkpoint_id,
                enables checkpointing for this run.
-            **kwargs: Additional keyword arguments.
+            **kwargs: Additional keyword arguments passed through to underlying workflow
+                and ai_function tools.

        Yields:
            AgentRunResponseUpdate objects representing the workflow execution progress.
@@ -198,7 +201,7 @@ class WorkflowAgent(BaseAgent):
        response_id = str(uuid.uuid4())

        async for update in self._run_stream_impl(
-            input_messages, response_id, thread, checkpoint_id, checkpoint_storage
+            input_messages, response_id, thread, checkpoint_id, checkpoint_storage, **kwargs
        ):
            response_updates.append(update)
            yield update
@@ -216,6 +219,7 @@ class WorkflowAgent(BaseAgent):
        thread: AgentThread,
        checkpoint_id: str | None = None,
        checkpoint_storage: CheckpointStorage | None = None,
+        **kwargs: Any,
    ) -> AsyncIterable[AgentRunResponseUpdate]:
        """Internal implementation of streaming execution.

@@ -225,6 +229,8 @@ class WorkflowAgent(BaseAgent):
            thread: The conversation thread containing message history.
            checkpoint_id: ID of checkpoint to restore from.
            checkpoint_storage: Runtime checkpoint storage.
+            **kwargs: Additional keyword arguments passed through to the underlying
+                workflow and ai_function tools.

        Yields:
            AgentRunResponseUpdate objects representing the workflow execution progress.
@@ -255,6 +261,7 @@ class WorkflowAgent(BaseAgent):
                message=None,
                checkpoint_id=checkpoint_id,
                checkpoint_storage=checkpoint_storage,
+                **kwargs,
            )
        else:
            # Execute workflow with streaming (initial run or no function responses)
@@ -268,6 +275,7 @@ class WorkflowAgent(BaseAgent):
            event_stream = self.workflow.run_stream(
                message=conversation_messages,
                checkpoint_storage=checkpoint_storage,
+                **kwargs,
            )

        # Process events from the stream
@@ -286,10 +294,20 @@ class WorkflowAgent(BaseAgent):

        AgentRunUpdateEvent, RequestInfoEvent, and WorkflowOutputEvent are processed.
        Other workflow events are ignored as they are workflow-internal.
+
+        For AgentRunUpdateEvent from AgentExecutor instances, only events from executors
+        with output_response=True are converted to agent updates. This prevents agent
+        responses from executors that were not explicitly marked to surface their output.
+        Non-AgentExecutor executors that emit AgentRunUpdateEvent directly are allowed
+        through since they explicitly chose to emit the event.
        """
        match event:
-            case AgentRunUpdateEvent(data=update):
-                # Direct pass-through of update in an agent streaming event
+            case AgentRunUpdateEvent(data=update, executor_id=executor_id):
+                # For AgentExecutor instances, only pass through if output_response=True.
+                # Non-AgentExecutor executors that emit AgentRunUpdateEvent are allowed through.
+                executor = self.workflow.executors.get(executor_id)
+                if isinstance(executor, AgentExecutor) and not executor.output_response:
+                    return None
                if update:
                    return update
                return None
@@ -297,11 +315,17 @@ class WorkflowAgent(BaseAgent):
            case WorkflowOutputEvent(data=data, source_executor_id=source_executor_id):
                # Convert workflow output to an agent response update.
                # Handle different data types appropriately.
+
+                # Skip AgentRunResponse from AgentExecutor with output_response=True
+                # since streaming events already surfaced the content.
+                if isinstance(data, AgentRunResponse):
+                    executor = self.workflow.executors.get(source_executor_id)
+                    if isinstance(executor, AgentExecutor) and executor.output_response:
+                        return None
+
                if isinstance(data, AgentRunResponseUpdate):
-                    # Already an update, pass through
                    return data
                if isinstance(data, ChatMessage):
-                    # Convert ChatMessage to update
                    return AgentRunResponseUpdate(
                        contents=list(data.contents),
                        role=data.role,
@@ -311,15 +335,9 @@ class WorkflowAgent(BaseAgent):
                        created_at=datetime.now(tz=timezone.utc).strftime("%Y-%m-%dT%H:%M:%S.%fZ"),
                        raw_representation=data,
                    )
-                # Determine contents based on data type
-                if isinstance(data, BaseContent):
-                    # Already a content type (TextContent, ImageContent, etc.)
-                    contents: list[Contents] = [cast(Contents, data)]
-                elif isinstance(data, str):
-                    contents = [TextContent(text=data)]
-                else:
-                    # Fallback: convert to string representation
-                    contents = [TextContent(text=str(data))]
+                contents = self._extract_contents(data)
+                if not contents:
+                    return None
                return AgentRunResponseUpdate(
                    contents=contents,
                    role=Role.ASSISTANT,
@@ -405,6 +423,18 @@ class WorkflowAgent(BaseAgent):
                        raise AgentExecutionException("Unexpected content type while awaiting request info responses.")
        return function_responses

+    def _extract_contents(self, data: Any) -> list[Contents]:
+        """Recursively extract Contents from workflow output data."""
+        if isinstance(data, ChatMessage):
+            return list(data.contents)
+        if isinstance(data, list):
+            return [c for item in data for c in self._extract_contents(item)]
+        if isinstance(data, BaseContent):
+            return [cast(Contents, data)]
+        if isinstance(data, str):
+            return [TextContent(text=data)]
+        return [TextContent(text=str(data))]
+
    class _ResponseState(TypedDict):
        """State for grouping response updates by message_id."""

@@ -99,6 +99,11 @@ class AgentExecutor(Executor):
        self._output_response = output_response
        self._cache: list[ChatMessage] = []

+    @property
+    def output_response(self) -> bool:
+        """Whether this executor yields AgentRunResponse as workflow output when complete."""
+        return self._output_response
+
    @property
    def workflow_output_types(self) -> list[type[Any]]:
        # Override to declare AgentRunResponse as a possible output type only if enabled.
@@ -871,8 +871,10 @@ class HandoffBuilder:
            HandoffBuilder(participants=[coordinator, refund, shipping])
            .set_coordinator(coordinator)
            .with_termination_condition(
-                lambda conv: sum(1 for msg in conv if msg.role.value == "user") >= 5
-                or any("goodbye" in msg.text.lower() for msg in conv[-2:])
+                lambda conv: (
+                    sum(1 for msg in conv if msg.role.value == "user") >= 5
+                    or any("goodbye" in msg.text.lower() for msg in conv[-2:])
+                )
            )
            .build()
        )
@@ -7,16 +7,16 @@ import uuid
 from pathlib import Path
 from typing import Literal

-from ._edge import FanInEdgeGroup
+from ._edge import FanInEdgeGroup, InternalEdgeGroup
 from ._workflow import Workflow

 # Import of WorkflowExecutor is performed lazily inside methods to avoid cycles

-"""Workflow visualization module using graphviz."""
+"""Workflow visualization module using graphviz and Mermaid."""


 class WorkflowViz:
-    """A class for visualizing workflows using graphviz."""
+    """A class for visualizing workflows using graphviz and Mermaid."""

    def __init__(self, workflow: Workflow):
        """Initialize the WorkflowViz with a workflow.
@@ -26,9 +26,13 @@ class WorkflowViz:
        """
        self._workflow = workflow

-    def to_digraph(self) -> str:
+    def to_digraph(self, include_internal_executors: bool = False) -> str:
        """Export the workflow as a DOT format digraph string.

+        Args:
+            include_internal_executors (bool): Whether to include internal executors in the visualization.
+                                               Default is False.
+
        Returns:
            A string representation of the workflow in DOT format.
        """
@@ -39,20 +43,37 @@ class WorkflowViz:
        lines.append("")

        # Emit the top-level workflow nodes/edges
-        self._emit_workflow_digraph(self._workflow, lines, indent="  ")
+        self._emit_workflow_digraph(
+            self._workflow,
+            lines,
+            indent="  ",
+            include_internal_executors=include_internal_executors,
+        )

        # Emit sub-workflows hosted by WorkflowExecutor as nested clusters
-        self._emit_sub_workflows_digraph(self._workflow, lines, indent="  ")
+        self._emit_sub_workflows_digraph(
+            self._workflow,
+            lines,
+            indent="  ",
+            include_internal_executors=include_internal_executors,
+        )

        lines.append("}")
        return "\n".join(lines)

-    def export(self, format: Literal["svg", "png", "pdf", "dot"] = "svg", filename: str | None = None) -> str:
+    def export(
+        self,
+        format: Literal["svg", "png", "pdf", "dot"] = "svg",
+        filename: str | None = None,
+        include_internal_executors: bool = False,
+    ) -> str:
        """Export the workflow visualization to a file or return the file path.

        Args:
            format: The output format. Supported formats: 'svg', 'png', 'pdf', 'dot'.
            filename: Optional filename to save the output. If None, creates a temporary file.
+            include_internal_executors (bool): Whether to include internal executors in the visualization.
+                                               Default is False.

        Returns:
            The path to the saved file.
@@ -66,7 +87,7 @@ class WorkflowViz:
            raise ValueError(f"Unsupported format: {format}. Supported formats: svg, png, pdf, dot")

        if format == "dot":
-            content = self.to_digraph()
+            content = self.to_digraph(include_internal_executors=include_internal_executors)
            if filename:
                with open(filename, "w", encoding="utf-8") as f:
                    f.write(content)
@@ -87,7 +108,7 @@ class WorkflowViz:
            ) from e

        # Create a temporary graphviz Source object
-        dot_content = self.to_digraph()
+        dot_content = self.to_digraph(include_internal_executors=include_internal_executors)
        source = graphviz.Source(dot_content)

        try:
@@ -99,7 +120,7 @@ class WorkflowViz:

                # Remove extension if present since graphviz.render() adds it
                base_name = str(output_path.with_suffix(""))
-                source.render(base_name, format=format, cleanup=True)
+                source.render(base_name, format=format, cleanup=True)  # type: ignore

                # Return the actual filename with extension
                return f"{base_name}.{format}"
@@ -108,7 +129,7 @@ class WorkflowViz:
                temp_path = Path(temp_file.name)
                base_name = str(temp_path.with_suffix(""))

-            source.render(base_name, format=format, cleanup=True)
+            source.render(base_name, format=format, cleanup=True)  # type: ignore
            return f"{base_name}.{format}"
        except graphviz.backend.execute.ExecutableNotFound as e:
            raise ImportError(
@@ -118,60 +139,72 @@ class WorkflowViz:
                "brew install graphviz on macOS, or download from https://graphviz.org/download/ for other platforms."
            ) from e

-    def save_svg(self, filename: str) -> str:
+    def save_svg(self, filename: str, include_internal_executors: bool = False) -> str:
        """Convenience method to save as SVG.

        Args:
            filename: The filename to save the SVG file.
+            include_internal_executors (bool): Whether to include internal executors in the visualization.
+                                               Default is False.

        Returns:
            The path to the saved SVG file.
        """
-        return self.export(format="svg", filename=filename)
+        return self.export(format="svg", filename=filename, include_internal_executors=include_internal_executors)

-    def save_png(self, filename: str) -> str:
+    def save_png(self, filename: str, include_internal_executors: bool = False) -> str:
        """Convenience method to save as PNG.

        Args:
            filename: The filename to save the PNG file.
+            include_internal_executors (bool): Whether to include internal executors in the visualization.
+                                               Default is False.

        Returns:
            The path to the saved PNG file.
        """
-        return self.export(format="png", filename=filename)
+        return self.export(format="png", filename=filename, include_internal_executors=include_internal_executors)

-    def save_pdf(self, filename: str) -> str:
+    def save_pdf(self, filename: str, include_internal_executors: bool = False) -> str:
        """Convenience method to save as PDF.

        Args:
            filename: The filename to save the PDF file.
+            include_internal_executors (bool): Whether to include internal executors in the visualization.
+                                               Default is False.

        Returns:
            The path to the saved PDF file.
        """
-        return self.export(format="pdf", filename=filename)
+        return self.export(format="pdf", filename=filename, include_internal_executors=include_internal_executors)

-    def to_mermaid(self) -> str:
+    def to_mermaid(self, include_internal_executors: bool = False) -> str:
        """Export the workflow as a Mermaid flowchart string.

+        Args:
+            include_internal_executors (bool): Whether to include internal executors in the visualization.
+                                               Default is False.
+
        Returns:
            A string representation of the workflow in Mermaid flowchart syntax.
        """
-
-        def _san(s: str) -> str:
-            """Sanitize an ID for Mermaid (alphanumeric and underscore, start with letter)."""
-            s2 = re.sub(r"[^0-9A-Za-z_]", "_", s)
-            if not s2 or not s2[0].isalpha():
-                s2 = f"n_{s2}"
-            return s2
-
        lines: list[str] = ["flowchart TD"]

        # Emit top-level workflow
-        self._emit_workflow_mermaid(self._workflow, lines, indent="  ")
+        self._emit_workflow_mermaid(
+            self._workflow,
+            lines,
+            indent="  ",
+            include_internal_executors=include_internal_executors,
+        )

        # Emit sub-workflows as Mermaid subgraphs
-        self._emit_sub_workflows_mermaid(self._workflow, lines, indent="  ")
+        self._emit_sub_workflows_mermaid(
+            self._workflow,
+            lines,
+            indent="  ",
+            include_internal_executors=include_internal_executors,
+        )

        return "\n".join(lines)

@@ -181,13 +214,13 @@ class WorkflowViz:
        sources_sorted = sorted(sources)
        return hashlib.sha256((target + "|" + "|".join(sources_sorted)).encode("utf-8")).hexdigest()[:8]

-    def _compute_fan_in_descriptors(self, wf: Workflow | None = None) -> list[tuple[str, list[str], str]]:
+    def _compute_fan_in_descriptors(self, workflow: Workflow | None = None) -> list[tuple[str, list[str], str]]:
        """Return list of (node_id, sources, target) for fan-in groups.

        node_id is DOT-oriented: fan_in::target::digest
        """
        result: list[tuple[str, list[str], str]] = []
-        workflow = wf or self._workflow
+        workflow = workflow or self._workflow
        for group in workflow.edge_groups:
            if isinstance(group, FanInEdgeGroup):
                target = group.target_executor_ids[0]
@@ -197,13 +230,19 @@ class WorkflowViz:
                result.append((node_id, sorted(sources), target))
        return result

-    def _compute_normal_edges(self, wf: Workflow | None = None) -> list[tuple[str, str, bool]]:
+    def _compute_normal_edges(
+        self,
+        workflow: Workflow | None = None,
+        include_internal_executors: bool = False,
+    ) -> list[tuple[str, str, bool]]:
        """Return list of (source_id, target_id, is_conditional) for non-fan-in groups."""
        edges: list[tuple[str, str, bool]] = []
-        workflow = wf or self._workflow
+        workflow = workflow or self._workflow
        for group in workflow.edge_groups:
            if isinstance(group, FanInEdgeGroup):
                continue
+            if isinstance(group, InternalEdgeGroup) and not include_internal_executors:
+                continue
            for edge in group.edges:
                is_cond = getattr(edge, "_condition", None) is not None
                edges.append((edge.source_id, edge.target_id, is_cond))
@@ -213,7 +252,14 @@ class WorkflowViz:

    # region Internal emitters (DOT)

-    def _emit_workflow_digraph(self, wf: Workflow, lines: list[str], indent: str, ns: str | None = None) -> None:
+    def _emit_workflow_digraph(
+        self,
+        workflow: Workflow,
+        lines: list[str],
+        indent: str,
+        ns: str | None = None,
+        include_internal_executors: bool = False,
+    ) -> None:
        """Emit DOT nodes/edges for the given workflow.

        If ns (namespace) is provided, node ids are prefixed with f"{ns}/" for uniqueness,
@@ -224,16 +270,16 @@ class WorkflowViz:
            return f"{ns}/{x}" if ns else x

        # Nodes
-        start_executor_id = wf.start_executor_id
+        start_executor_id = workflow.start_executor_id
        lines.append(
            f'{indent}"{map_id(start_executor_id)}" [fillcolor=lightgreen, label="{start_executor_id}\\n(Start)"];'
        )
-        for executor_id in wf.executors:
+        for executor_id in workflow.executors:
            if executor_id != start_executor_id:
                lines.append(f'{indent}"{map_id(executor_id)}" [label="{executor_id}"];')

        # Fan-in nodes
-        fan_in_nodes = self._compute_fan_in_descriptors(wf)
+        fan_in_nodes = self._compute_fan_in_descriptors(workflow)
        if fan_in_nodes:
            lines.append("")
            for node_id, _, _ in fan_in_nodes:
@@ -246,11 +292,19 @@ class WorkflowViz:
            lines.append(f'{indent}"{map_id(node_id)}" -> "{map_id(target)}";')

        # Normal edges
-        for src, tgt, is_cond in self._compute_normal_edges(wf):
+        for src, tgt, is_cond in self._compute_normal_edges(
+            workflow, include_internal_executors=include_internal_executors
+        ):
            edge_attr = ' [style=dashed, label="conditional"]' if is_cond else ""
            lines.append(f'{indent}"{map_id(src)}" -> "{map_id(tgt)}"{edge_attr};')

-    def _emit_sub_workflows_digraph(self, wf: Workflow, lines: list[str], indent: str) -> None:
+    def _emit_sub_workflows_digraph(
+        self,
+        workflow: Workflow,
+        lines: list[str],
+        indent: str,
+        include_internal_executors: bool = False,
+    ) -> None:
        """Emit DOT subgraphs for any WorkflowExecutor instances found in the workflow."""
        # Lazy import to avoid any potential import cycles
        try:
@@ -258,7 +312,7 @@ class WorkflowViz:
        except ImportError:  # pragma: no cover - best-effort; if unavailable, skip subgraphs
            return

-        for exec_id, exec_obj in wf.executors.items():
+        for exec_id, exec_obj in workflow.executors.items():
            if isinstance(exec_obj, WorkflowExecutor) and hasattr(exec_obj, "workflow") and exec_obj.workflow:
                subgraph_id = f"cluster_{uuid.uuid5(uuid.NAMESPACE_OID, exec_id).hex[:8]}"
                lines.append(f"{indent}subgraph {subgraph_id} {{")
@@ -267,10 +321,21 @@ class WorkflowViz:

                # Emit the nested workflow inside this cluster using a namespace
                ns = exec_id
-                self._emit_workflow_digraph(exec_obj.workflow, lines, indent=f"{indent}  ", ns=ns)
+                self._emit_workflow_digraph(
+                    exec_obj.workflow,
+                    lines,
+                    indent=f"{indent}  ",
+                    ns=ns,
+                    include_internal_executors=include_internal_executors,
+                )

                # Recurse into deeper nested sub-workflows
-                self._emit_sub_workflows_digraph(exec_obj.workflow, lines, indent=f"{indent}  ")
+                self._emit_sub_workflows_digraph(
+                    exec_obj.workflow,
+                    lines,
+                    indent=f"{indent}  ",
+                    include_internal_executors=include_internal_executors,
+                )

                lines.append(f"{indent}}}")

@@ -278,7 +343,14 @@ class WorkflowViz:

    # region Internal emitters (Mermaid)

-    def _emit_workflow_mermaid(self, wf: Workflow, lines: list[str], indent: str, ns: str | None = None) -> None:
+    def _emit_workflow_mermaid(
+        self,
+        workflow: Workflow,
+        lines: list[str],
+        indent: str,
+        ns: str | None = None,
+        include_internal_executors: bool = False,
+    ) -> None:
        def _san(s: str) -> str:
            s2 = re.sub(r"[^0-9A-Za-z_]", "_", s)
            if not s2 or not s2[0].isalpha():
@@ -291,15 +363,15 @@ class WorkflowViz:
            return _san(x)

        # Nodes
-        start_executor_id = wf.start_executor_id
+        start_executor_id = workflow.start_executor_id
        lines.append(f'{indent}{map_id(start_executor_id)}["{start_executor_id} (Start)"];')
-        for executor_id in wf.executors:
+        for executor_id in workflow.executors:
            if executor_id == start_executor_id:
                continue
            lines.append(f'{indent}{map_id(executor_id)}["{executor_id}"];')

        # Fan-in nodes
-        fan_in_nodes_dot = self._compute_fan_in_descriptors(wf)
+        fan_in_nodes_dot = self._compute_fan_in_descriptors(workflow)
        fan_in_nodes: list[tuple[str, list[str], str]] = []
        for dot_node_id, sources, target in fan_in_nodes_dot:
            digest = dot_node_id.split("::")[-1]
@@ -318,7 +390,9 @@ class WorkflowViz:
            lines.append(f"{indent}{fan_node_id} --> {map_id(target)};")

        # Normal edges
-        for src, tgt, is_cond in self._compute_normal_edges(wf):
+        for src, tgt, is_cond in self._compute_normal_edges(
+            workflow, include_internal_executors=include_internal_executors
+        ):
            s = map_id(src)
            t = map_id(tgt)
            if is_cond:
@@ -326,7 +400,13 @@ class WorkflowViz:
            else:
                lines.append(f"{indent}{s} --> {t};")

-    def _emit_sub_workflows_mermaid(self, wf: Workflow, lines: list[str], indent: str) -> None:
+    def _emit_sub_workflows_mermaid(
+        self,
+        workflow: Workflow,
+        lines: list[str],
+        indent: str,
+        include_internal_executors: bool = False,
+    ) -> None:
        try:
            from ._workflow_executor import WorkflowExecutor  # type: ignore
        except ImportError:  # pragma: no cover
@@ -338,14 +418,25 @@ class WorkflowViz:
                s2 = f"n_{s2}"
            return s2

-        for exec_id, exec_obj in wf.executors.items():
+        for exec_id, exec_obj in workflow.executors.items():
            if isinstance(exec_obj, WorkflowExecutor) and hasattr(exec_obj, "workflow") and exec_obj.workflow:
                sg_id = _san(exec_id)
                lines.append(f"{indent}subgraph {sg_id}")
                # Render nested workflow within this subgraph using namespacing
-                self._emit_workflow_mermaid(exec_obj.workflow, lines, indent=f"{indent}  ", ns=exec_id)
+                self._emit_workflow_mermaid(
+                    exec_obj.workflow,
+                    lines,
+                    indent=f"{indent}  ",
+                    ns=exec_id,
+                    include_internal_executors=include_internal_executors,
+                )
                # Recurse into deeper sub-workflows
-                self._emit_sub_workflows_mermaid(exec_obj.workflow, lines, indent=f"{indent}  ")
+                self._emit_sub_workflows_mermaid(
+                    exec_obj.workflow,
+                    lines,
+                    indent=f"{indent}  ",
+                    include_internal_executors=include_internal_executors,
+                )
                lines.append(f"{indent}end")

    # endregion
@@ -11,6 +11,7 @@ if TYPE_CHECKING:
    from ._workflow import Workflow

 from ._checkpoint_encoding import decode_checkpoint_value, encode_checkpoint_value
+from ._const import WORKFLOW_RUN_KWARGS_KEY
 from ._events import (
    RequestInfoEvent,
    WorkflowErrorEvent,
@@ -366,8 +367,11 @@ class WorkflowExecutor(Executor):
        logger.debug(f"WorkflowExecutor {self.id} starting sub-workflow {self.workflow.id} execution {execution_id}")

        try:
-            # Run the sub-workflow and collect all events
-            result = await self.workflow.run(input_data)
+            # Get kwargs from parent workflow's SharedState to propagate to subworkflow
+            parent_kwargs: dict[str, Any] = await ctx.get_shared_state(WORKFLOW_RUN_KWARGS_KEY) or {}
+
+            # Run the sub-workflow and collect all events, passing parent kwargs
+            result = await self.workflow.run(input_data, **parent_kwargs)

            logger.debug(
                f"WorkflowExecutor {self.id} sub-workflow {self.workflow.id} "
@@ -154,7 +154,7 @@ class AzureOpenAIChatClient(AzureOpenAIConfigMixin, OpenAIBaseChatClient):
        )

    @override
-    def _parse_text_from_choice(self, choice: Choice | ChunkChoice) -> TextContent | None:
+    def _parse_text_from_openai(self, choice: Choice | ChunkChoice) -> TextContent | None:
        """Parse the choice into a TextContent object.

        Overwritten from OpenAIBaseChatClient to deal with Azure On Your Data function.
@@ -1680,13 +1680,12 @@ def _capture_messages(
    prepped = prepare_messages(messages, system_instructions=system_instructions)
    otel_messages: list[dict[str, Any]] = []
    for index, message in enumerate(prepped):
-        otel_messages.append(_to_otel_message(message))
-        try:
-            message_data = message.to_dict(exclude_none=True)
-        except Exception:
-            message_data = {"role": message.role.value, "contents": message.contents}
+        # Reuse the otel message representation for logging instead of calling to_dict()
+        # to avoid expensive Pydantic serialization overhead
+        otel_message = _to_otel_message(message)
+        otel_messages.append(otel_message)
        logger.info(
-            message_data,
+            otel_message,
            extra={
                OtelAttr.EVENT_NAME: OtelAttr.CHOICE if output else ROLE_EVENT_MAP.get(message.role.value),
                OtelAttr.PROVIDER_NAME: provider_name,
@@ -0,0 +1,23 @@
+# Copyright (c) Microsoft. All rights reserved.
+
+import importlib
+from typing import Any
+
+IMPORT_PATH = "agent_framework_ollama"
+PACKAGE_NAME = "agent-framework-ollama"
+_IMPORTS = ["__version__", "OllamaChatClient", "OllamaSettings"]
+
+
+def __getattr__(name: str) -> Any:
+    if name in _IMPORTS:
+        try:
+            return getattr(importlib.import_module(IMPORT_PATH), name)
+        except ModuleNotFoundError as exc:
+            raise ModuleNotFoundError(
+                f"The '{PACKAGE_NAME}' package is not installed, please do `pip install {PACKAGE_NAME}`"
+            ) from exc
+    raise AttributeError(f"Module {IMPORT_PATH} has no attribute {name}.")
+
+
+def __dir__() -> list[str]:
+    return _IMPORTS
@@ -0,0 +1,13 @@
+# Copyright (c) Microsoft. All rights reserved.
+
+from agent_framework_ollama import (
+    OllamaChatClient,
+    OllamaSettings,
+    __version__,
+)
+
+__all__ = [
+    "OllamaChatClient",
+    "OllamaSettings",
+    "__version__",
+]
@@ -164,7 +164,7 @@ class OpenAIAssistantsClient(OpenAIConfigMixin, BaseChatClient):
    async def close(self) -> None:
        """Clean up any assistants we created."""
        if self._should_delete_assistant and self.assistant_id is not None:
-            client = await self.ensure_client()
+            client = await self._ensure_client()
            await client.beta.assistants.delete(self.assistant_id)
            object.__setattr__(self, "assistant_id", None)
            object.__setattr__(self, "_should_delete_assistant", False)
@@ -188,7 +188,7 @@ class OpenAIAssistantsClient(OpenAIConfigMixin, BaseChatClient):
        chat_options: ChatOptions,
        **kwargs: Any,
    ) -> AsyncIterable[ChatResponseUpdate]:
-        # Extract necessary state from messages and options
+        # prepare
        run_options, tool_results = self._prepare_options(messages, chat_options, **kwargs)

        # Get the thread ID
@@ -204,10 +204,10 @@ class OpenAIAssistantsClient(OpenAIConfigMixin, BaseChatClient):
        # Determine which assistant to use and create if needed
        assistant_id = await self._get_assistant_id_or_create()

-        # Create the streaming response
+        # execute
        stream, thread_id = await self._create_assistant_stream(thread_id, assistant_id, run_options, tool_results)

-        # Process and yield each update from the stream
+        # process
        async for update in self._process_stream_events(stream, thread_id):
            yield update

@@ -222,7 +222,7 @@ class OpenAIAssistantsClient(OpenAIConfigMixin, BaseChatClient):
            if not self.model_id:
                raise ServiceInitializationError("Parameter 'model_id' is required for assistant creation.")

-            client = await self.ensure_client()
+            client = await self._ensure_client()
            created_assistant = await client.beta.assistants.create(
                model=self.model_id,
                description=self.assistant_description,
@@ -245,11 +245,11 @@ class OpenAIAssistantsClient(OpenAIConfigMixin, BaseChatClient):
        Returns:
            tuple: (stream, final_thread_id)
        """
-        client = await self.ensure_client()
+        client = await self._ensure_client()
        # Get any active run for this thread
        thread_run = await self._get_active_thread_run(thread_id)

-        tool_run_id, tool_outputs = self._convert_function_results_to_tool_output(tool_results)
+        tool_run_id, tool_outputs = self._prepare_tool_outputs_for_assistants(tool_results)

        if thread_run is not None and tool_run_id is not None and tool_run_id == thread_run.id and tool_outputs:
            # There's an active run and we have tool results to submit, so submit the results.
@@ -270,7 +270,7 @@ class OpenAIAssistantsClient(OpenAIConfigMixin, BaseChatClient):

    async def _get_active_thread_run(self, thread_id: str | None) -> Run | None:
        """Get any active run for the given thread."""
-        client = await self.ensure_client()
+        client = await self._ensure_client()
        if thread_id is None:
            return None

@@ -281,7 +281,7 @@ class OpenAIAssistantsClient(OpenAIConfigMixin, BaseChatClient):

    async def _prepare_thread(self, thread_id: str | None, thread_run: Run | None, run_options: dict[str, Any]) -> str:
        """Prepare the thread for a new run, creating or cleaning up as needed."""
-        client = await self.ensure_client()
+        client = await self._ensure_client()
        if thread_id is None:
            # No thread ID was provided, so create a new thread.
            thread = await client.beta.threads.create(  # type: ignore[reportDeprecated]
@@ -330,7 +330,7 @@ class OpenAIAssistantsClient(OpenAIConfigMixin, BaseChatClient):
                                response_id=response_id,
                            )
                elif response.event == "thread.run.requires_action" and isinstance(response.data, Run):
-                    contents = self._create_function_call_contents(response.data, response_id)
+                    contents = self._parse_function_calls_from_assistants(response.data, response_id)
                    if contents:
                        yield ChatResponseUpdate(
                            role=Role.ASSISTANT,
@@ -371,8 +371,8 @@ class OpenAIAssistantsClient(OpenAIConfigMixin, BaseChatClient):
                        role=Role.ASSISTANT,
                    )

-    def _create_function_call_contents(self, event_data: Run, response_id: str | None) -> list[Contents]:
-        """Create function call contents from a tool action event."""
+    def _parse_function_calls_from_assistants(self, event_data: Run, response_id: str | None) -> list[Contents]:
+        """Parse function call contents from an assistants tool action event."""
        contents: list[Contents] = []

        if event_data.required_action is not None:
@@ -437,7 +437,10 @@ class OpenAIAssistantsClient(OpenAIConfigMixin, BaseChatClient):
            if chat_options.response_format is not None:
                run_options["response_format"] = {
                    "type": "json_schema",
-                    "json_schema": chat_options.response_format.model_json_schema(),
+                    "json_schema": {
+                        "name": chat_options.response_format.__name__,
+                        "schema": chat_options.response_format.model_json_schema(),
+                    },
                }

        instructions: list[str] = []
@@ -487,10 +490,11 @@ class OpenAIAssistantsClient(OpenAIConfigMixin, BaseChatClient):

        return run_options, tool_results

-    def _convert_function_results_to_tool_output(
+    def _prepare_tool_outputs_for_assistants(
        self,
        tool_results: list[FunctionResultContent] | None,
    ) -> tuple[str | None, list[ToolOutput] | None]:
+        """Prepare function results for submission to the assistants API."""
        run_id: str | None = None
        tool_outputs: list[ToolOutput] | None = None

@@ -14,7 +14,7 @@ from openai.types.chat.chat_completion import ChatCompletion, Choice
 from openai.types.chat.chat_completion_chunk import ChatCompletionChunk
 from openai.types.chat.chat_completion_chunk import Choice as ChunkChoice
 from openai.types.chat.chat_completion_message_custom_tool_call import ChatCompletionMessageCustomToolCall
-from pydantic import BaseModel, ValidationError
+from pydantic import ValidationError

 from .._clients import BaseChatClient
 from .._logging import get_logger
@@ -34,6 +34,7 @@ from .._types import (
    FunctionResultContent,
    Role,
    TextContent,
+    TextReasoningContent,
    UriContent,
    UsageContent,
    UsageDetails,
@@ -69,10 +70,12 @@ class OpenAIBaseChatClient(OpenAIBase, BaseChatClient):
        chat_options: ChatOptions,
        **kwargs: Any,
    ) -> ChatResponse:
-        client = await self.ensure_client()
+        client = await self._ensure_client()
+        # prepare
        options_dict = self._prepare_options(messages, chat_options)
        try:
-            return self._create_chat_response(
+            # execute and process
+            return self._parse_response_from_openai(
                await client.chat.completions.create(stream=False, **options_dict), chat_options
            )
        except BadRequestError as ex:
@@ -98,14 +101,16 @@ class OpenAIBaseChatClient(OpenAIBase, BaseChatClient):
        chat_options: ChatOptions,
        **kwargs: Any,
    ) -> AsyncIterable[ChatResponseUpdate]:
-        client = await self.ensure_client()
+        client = await self._ensure_client()
+        # prepare
        options_dict = self._prepare_options(messages, chat_options)
        options_dict["stream_options"] = {"include_usage": True}
        try:
+            # execute and process
            async for chunk in await client.chat.completions.create(stream=True, **options_dict):
                if len(chunk.choices) == 0 and chunk.usage is None:
                    continue
-                yield self._create_chat_response_update(chunk)
+                yield self._parse_response_update_from_openai(chunk)
        except BadRequestError as ex:
            if ex.code == "content_filter":
                raise OpenAIContentFilterException(
@@ -124,7 +129,9 @@ class OpenAIBaseChatClient(OpenAIBase, BaseChatClient):

    # region content creation

-    def _chat_to_tool_spec(self, tools: Sequence[ToolProtocol | MutableMapping[str, Any]]) -> list[dict[str, Any]]:
+    def _prepare_tools_for_openai(
+        self, tools: Sequence[ToolProtocol | MutableMapping[str, Any]]
+    ) -> list[dict[str, Any]]:
        chat_tools: list[dict[str, Any]] = []
        for tool in tools:
            if isinstance(tool, ToolProtocol):
@@ -157,51 +164,65 @@ class OpenAIBaseChatClient(OpenAIBase, BaseChatClient):
        return None

    def _prepare_options(self, messages: MutableSequence[ChatMessage], chat_options: ChatOptions) -> dict[str, Any]:
-        # Preprocess web search tool if it exists
-        options_dict = chat_options.to_dict(
+        run_options = chat_options.to_dict(
            exclude={
                "type",
                "instructions",  # included as system message
-                "allow_multiple_tool_calls",  # handled separately
+                "response_format",  # handled separately
+                "additional_properties",  # handled separately
            }
        )

-        if messages and "messages" not in options_dict:
-            options_dict["messages"] = self._prepare_chat_history_for_request(messages)
-        if "messages" not in options_dict:
+        # messages
+        if messages and "messages" not in run_options:
+            run_options["messages"] = self._prepare_messages_for_openai(messages)
+        if "messages" not in run_options:
            raise ServiceInvalidRequestError("Messages are required for chat completions")
+
+        # Translation between ChatOptions and Chat Completion API
+        translations = {
+            "model_id": "model",
+            "allow_multiple_tool_calls": "parallel_tool_calls",
+            "max_tokens": "max_output_tokens",
+        }
+        for old_key, new_key in translations.items():
+            if old_key in run_options and old_key != new_key:
+                run_options[new_key] = run_options.pop(old_key)
+
+        # model id
+        if not run_options.get("model"):
+            if not self.model_id:
+                raise ValueError("model_id must be a non-empty string")
+            run_options["model"] = self.model_id
+
+        # tools
        if chat_options.tools is not None:
-            web_search_options = self._process_web_search_tool(chat_options.tools)
-            if web_search_options:
-                options_dict["web_search_options"] = web_search_options
-            options_dict["tools"] = self._chat_to_tool_spec(chat_options.tools)
-        if chat_options.allow_multiple_tool_calls is not None:
-            options_dict["parallel_tool_calls"] = chat_options.allow_multiple_tool_calls
-        if not options_dict.get("tools", None):
-            options_dict.pop("tools", None)
-            options_dict.pop("parallel_tool_calls", None)
-            options_dict.pop("tool_choice", None)
+            # Preprocess web search tool if it exists
+            if web_search_options := self._process_web_search_tool(chat_options.tools):
+                run_options["web_search_options"] = web_search_options
+            run_options["tools"] = self._prepare_tools_for_openai(chat_options.tools)
+        if not run_options.get("tools", None):
+            run_options.pop("tools", None)
+            run_options.pop("parallel_tool_calls", None)
+            run_options.pop("tool_choice", None)
+        # tool choice when `tool_choice` is a dict with single key `mode`, extract the mode value
+        if (tool_choice := run_options.get("tool_choice")) and len(tool_choice.keys()) == 1:
+            run_options["tool_choice"] = tool_choice["mode"]

-        if "model_id" not in options_dict:
-            options_dict["model"] = self.model_id
-        else:
-            options_dict["model"] = options_dict.pop("model_id")
-        if (
-            chat_options.response_format
-            and isinstance(chat_options.response_format, type)
-            and issubclass(chat_options.response_format, BaseModel)
-        ):
-            options_dict["response_format"] = type_to_response_format_param(chat_options.response_format)
-        if additional_properties := options_dict.pop("additional_properties", None):
-            for key, value in additional_properties.items():
-                if value is not None:
-                    options_dict[key] = value
-        if (tool_choice := options_dict.get("tool_choice")) and len(tool_choice.keys()) == 1:
-            options_dict["tool_choice"] = tool_choice["mode"]
-        return options_dict
+        # response format
+        if chat_options.response_format:
+            run_options["response_format"] = type_to_response_format_param(chat_options.response_format)

-    def _create_chat_response(self, response: ChatCompletion, chat_options: ChatOptions) -> "ChatResponse":
-        """Create a chat message content object from a choice."""
+        # additional properties
+        additional_options = {
+            key: value for key, value in chat_options.additional_properties.items() if value is not None
+        }
+        if additional_options:
+            run_options.update(additional_options)
+        return run_options
+
+    def _parse_response_from_openai(self, response: ChatCompletion, chat_options: ChatOptions) -> "ChatResponse":
+        """Parse a response from OpenAI into a ChatResponse."""
        response_metadata = self._get_metadata_from_chat_response(response)
        messages: list[ChatMessage] = []
        finish_reason: FinishReason | None = None
@@ -210,15 +231,17 @@ class OpenAIBaseChatClient(OpenAIBase, BaseChatClient):
            if choice.finish_reason:
                finish_reason = FinishReason(value=choice.finish_reason)
            contents: list[Contents] = []
-            if text_content := self._parse_text_from_choice(choice):
+            if text_content := self._parse_text_from_openai(choice):
                contents.append(text_content)
-            if parsed_tool_calls := [tool for tool in self._get_tool_calls_from_chat_choice(choice)]:
+            if parsed_tool_calls := [tool for tool in self._parse_tool_calls_from_openai(choice)]:
                contents.extend(parsed_tool_calls)
+            if reasoning_details := getattr(choice.message, "reasoning_details", None):
+                contents.append(TextReasoningContent(None, protected_data=json.dumps(reasoning_details)))
            messages.append(ChatMessage(role="assistant", contents=contents))
        return ChatResponse(
            response_id=response.id,
            created_at=datetime.fromtimestamp(response.created, tz=timezone.utc).strftime("%Y-%m-%dT%H:%M:%S.%fZ"),
-            usage_details=self._usage_details_from_openai(response.usage) if response.usage else None,
+            usage_details=self._parse_usage_from_openai(response.usage) if response.usage else None,
            messages=messages,
            model_id=response.model,
            additional_properties=response_metadata,
@@ -226,16 +249,16 @@ class OpenAIBaseChatClient(OpenAIBase, BaseChatClient):
            response_format=chat_options.response_format,
        )

-    def _create_chat_response_update(
+    def _parse_response_update_from_openai(
        self,
        chunk: ChatCompletionChunk,
    ) -> ChatResponseUpdate:
-        """Create a streaming chat message content object from a choice."""
+        """Parse a streaming response update from OpenAI."""
        chunk_metadata = self._get_metadata_from_streaming_chat_response(chunk)
        if chunk.usage:
            return ChatResponseUpdate(
                role=Role.ASSISTANT,
-                contents=[UsageContent(details=self._usage_details_from_openai(chunk.usage), raw_representation=chunk)],
+                contents=[UsageContent(details=self._parse_usage_from_openai(chunk.usage), raw_representation=chunk)],
                model_id=chunk.model,
                additional_properties=chunk_metadata,
                response_id=chunk.id,
@@ -245,12 +268,14 @@ class OpenAIBaseChatClient(OpenAIBase, BaseChatClient):
        finish_reason: FinishReason | None = None
        for choice in chunk.choices:
            chunk_metadata.update(self._get_metadata_from_chat_choice(choice))
-            contents.extend(self._get_tool_calls_from_chat_choice(choice))
+            contents.extend(self._parse_tool_calls_from_openai(choice))
            if choice.finish_reason:
                finish_reason = FinishReason(value=choice.finish_reason)

-            if text_content := self._parse_text_from_choice(choice):
+            if text_content := self._parse_text_from_openai(choice):
                contents.append(text_content)
+            if reasoning_details := getattr(choice.delta, "reasoning_details", None):
+                contents.append(TextReasoningContent(None, protected_data=json.dumps(reasoning_details)))
        return ChatResponseUpdate(
            created_at=datetime.fromtimestamp(chunk.created, tz=timezone.utc).strftime("%Y-%m-%dT%H:%M:%S.%fZ"),
            contents=contents,
@@ -263,7 +288,7 @@ class OpenAIBaseChatClient(OpenAIBase, BaseChatClient):
            message_id=chunk.id,
        )

-    def _usage_details_from_openai(self, usage: CompletionUsage) -> UsageDetails:
+    def _parse_usage_from_openai(self, usage: CompletionUsage) -> UsageDetails:
        details = UsageDetails(
            input_token_count=usage.prompt_tokens,
            output_token_count=usage.completion_tokens,
@@ -285,7 +310,7 @@ class OpenAIBaseChatClient(OpenAIBase, BaseChatClient):
                details["prompt/cached_tokens"] = tokens
        return details

-    def _parse_text_from_choice(self, choice: Choice | ChunkChoice) -> TextContent | None:
+    def _parse_text_from_openai(self, choice: Choice | ChunkChoice) -> TextContent | None:
        """Parse the choice into a TextContent object."""
        message = choice.message if isinstance(choice, Choice) else choice.delta
        if message.content:
@@ -312,8 +337,8 @@ class OpenAIBaseChatClient(OpenAIBase, BaseChatClient):
            "logprobs": getattr(choice, "logprobs", None),
        }

-    def _get_tool_calls_from_chat_choice(self, choice: Choice | ChunkChoice) -> list[Contents]:
-        """Get tool calls from a chat choice."""
+    def _parse_tool_calls_from_openai(self, choice: Choice | ChunkChoice) -> list[Contents]:
+        """Parse tool calls from an OpenAI response choice."""
        resp: list[Contents] = []
        content = choice.message if isinstance(choice, Choice) else choice.delta
        if content and content.tool_calls:
@@ -331,13 +356,13 @@ class OpenAIBaseChatClient(OpenAIBase, BaseChatClient):
        # When you enable asynchronous content filtering in Azure OpenAI, you may receive empty deltas
        return resp

-    def _prepare_chat_history_for_request(
+    def _prepare_messages_for_openai(
        self,
        chat_messages: Sequence[ChatMessage],
        role_key: str = "role",
        content_key: str = "content",
    ) -> list[dict[str, Any]]:
-        """Prepare the chat history for a request.
+        """Prepare the chat history for an OpenAI request.

        Allowing customization of the key names for role/author, and optionally overriding the role.

@@ -355,14 +380,14 @@ class OpenAIBaseChatClient(OpenAIBase, BaseChatClient):
        Returns:
            prepared_chat_history (Any): The prepared chat history for a request.
        """
-        list_of_list = [self._openai_chat_message_parser(message) for message in chat_messages]
+        list_of_list = [self._prepare_message_for_openai(message) for message in chat_messages]
        # Flatten the list of lists into a single list
        return list(chain.from_iterable(list_of_list))

    # region Parsers

-    def _openai_chat_message_parser(self, message: ChatMessage) -> list[dict[str, Any]]:
-        """Parse a chat message into the openai format."""
+    def _prepare_message_for_openai(self, message: ChatMessage) -> list[dict[str, Any]]:
+        """Prepare a chat message for OpenAI."""
        all_messages: list[dict[str, Any]] = []
        for content in message.contents:
            # Skip approval content - it's internal framework state, not for the LLM
@@ -372,28 +397,36 @@ class OpenAIBaseChatClient(OpenAIBase, BaseChatClient):
            args: dict[str, Any] = {
                "role": message.role.value if isinstance(message.role, Role) else message.role,
            }
+            if message.author_name and message.role != Role.TOOL:
+                args["name"] = message.author_name
+            if "reasoning_details" in message.additional_properties and (
+                details := message.additional_properties["reasoning_details"]
+            ):
+                args["reasoning_details"] = details
            match content:
                case FunctionCallContent():
                    if all_messages and "tool_calls" in all_messages[-1]:
                        # If the last message already has tool calls, append to it
-                        all_messages[-1]["tool_calls"].append(self._openai_content_parser(content))
+                        all_messages[-1]["tool_calls"].append(self._prepare_content_for_openai(content))
                    else:
-                        args["tool_calls"] = [self._openai_content_parser(content)]  # type: ignore
+                        args["tool_calls"] = [self._prepare_content_for_openai(content)]  # type: ignore
                case FunctionResultContent():
                    args["tool_call_id"] = content.call_id
                    if content.result is not None:
                        args["content"] = prepare_function_call_results(content.result)
+                case TextReasoningContent(protected_data=protected_data) if protected_data is not None:
+                    all_messages[-1]["reasoning_details"] = json.loads(protected_data)
                case _:
                    if "content" not in args:
                        args["content"] = []
                    # this is a list to allow multi-modal content
-                    args["content"].append(self._openai_content_parser(content))  # type: ignore
+                    args["content"].append(self._prepare_content_for_openai(content))  # type: ignore
            if "content" in args or "tool_calls" in args:
                all_messages.append(args)
        return all_messages

-    def _openai_content_parser(self, content: Contents) -> dict[str, Any]:
-        """Parse contents into the openai format."""
+    def _prepare_content_for_openai(self, content: Contents) -> dict[str, Any]:
+        """Prepare content for OpenAI."""
        match content:
            case FunctionCallContent():
                args = json.dumps(content.arguments) if isinstance(content.arguments, Mapping) else content.arguments
@@ -89,28 +89,16 @@ class OpenAIBaseResponsesClient(OpenAIBase, BaseChatClient):
        chat_options: ChatOptions,
        **kwargs: Any,
    ) -> ChatResponse:
-        client = await self.ensure_client()
-        run_options = await self.prepare_options(messages, chat_options, **kwargs)
-        response_format = run_options.pop("response_format", None)
-        text_config = run_options.pop("text", None)
-        text_format, text_config = self._prepare_text_config(response_format=response_format, text_config=text_config)
-        if text_config:
-            run_options["text"] = text_config
+        client = await self._ensure_client()
+        # prepare
+        run_options = await self._prepare_options(messages, chat_options, **kwargs)
        try:
-            if not text_format:
-                response = await client.responses.create(
-                    stream=False,
-                    **run_options,
-                )
-                chat_options.conversation_id = self.get_conversation_id(response, chat_options.store)
-                return self._create_response_content(response, chat_options=chat_options)
-            parsed_response: ParsedResponse[BaseModel] = await client.responses.parse(
-                text_format=text_format,
-                stream=False,
-                **run_options,
-            )
-            chat_options.conversation_id = self.get_conversation_id(parsed_response, chat_options.store)
-            return self._create_response_content(parsed_response, chat_options=chat_options)
+            # execute and process
+            if "text_format" in run_options:
+                response = await client.responses.parse(stream=False, **run_options)
+            else:
+                response = await client.responses.create(stream=False, **run_options)
+            return self._parse_response_from_openai(response, chat_options=chat_options)
        except BadRequestError as ex:
            if ex.code == "content_filter":
                raise OpenAIContentFilterException(
@@ -134,35 +122,23 @@ class OpenAIBaseResponsesClient(OpenAIBase, BaseChatClient):
        chat_options: ChatOptions,
        **kwargs: Any,
    ) -> AsyncIterable[ChatResponseUpdate]:
-        client = await self.ensure_client()
-        run_options = await self.prepare_options(messages, chat_options, **kwargs)
+        client = await self._ensure_client()
+        # prepare
+        run_options = await self._prepare_options(messages, chat_options, **kwargs)
        function_call_ids: dict[int, tuple[str, str]] = {}  # output_index: (call_id, name)
-        response_format = run_options.pop("response_format", None)
-        text_config = run_options.pop("text", None)
-        text_format, text_config = self._prepare_text_config(response_format=response_format, text_config=text_config)
-        if text_config:
-            run_options["text"] = text_config
        try:
-            if not text_format:
-                response = await client.responses.create(
-                    stream=True,
-                    **run_options,
-                )
-                async for chunk in response:
-                    update = self._create_streaming_response_content(
+            # execute and process
+            if "text_format" not in run_options:
+                async for chunk in await client.responses.create(stream=True, **run_options):
+                    yield self._parse_chunk_from_openai(
                        chunk, chat_options=chat_options, function_call_ids=function_call_ids
                    )
-                    yield update
                return
-            async with client.responses.stream(
-                text_format=text_format,
-                **run_options,
-            ) as response:
+            async with client.responses.stream(**run_options) as response:
                async for chunk in response:
-                    update = self._create_streaming_response_content(
+                    yield self._parse_chunk_from_openai(
                        chunk, chat_options=chat_options, function_call_ids=function_call_ids
                    )
-                    yield update
        except BadRequestError as ex:
            if ex.code == "content_filter":
                raise OpenAIContentFilterException(
@@ -179,33 +155,33 @@ class OpenAIBaseResponsesClient(OpenAIBase, BaseChatClient):
                inner_exception=ex,
            ) from ex

-    def _prepare_text_config(
+    def _prepare_response_and_text_format(
        self,
        *,
        response_format: Any,
        text_config: MutableMapping[str, Any] | None,
    ) -> tuple[type[BaseModel] | None, dict[str, Any] | None]:
        """Normalize response_format into Responses text configuration and parse target."""
-        prepared_text = dict(text_config) if isinstance(text_config, MutableMapping) else None
        if text_config is not None and not isinstance(text_config, MutableMapping):
            raise ServiceInvalidRequestError("text must be a mapping when provided.")
+        text_config = cast(dict[str, Any], text_config) if isinstance(text_config, MutableMapping) else None

        if response_format is None:
-            return None, prepared_text
+            return None, text_config

        if isinstance(response_format, type) and issubclass(response_format, BaseModel):
-            if prepared_text and "format" in prepared_text:
+            if text_config and "format" in text_config:
                raise ServiceInvalidRequestError("response_format cannot be combined with explicit text.format.")
-            return response_format, prepared_text
+            return response_format, text_config

        if isinstance(response_format, Mapping):
            format_config = self._convert_response_format(cast("Mapping[str, Any]", response_format))
-            if prepared_text is None:
-                prepared_text = {}
-            elif "format" in prepared_text and prepared_text["format"] != format_config:
+            if text_config is None:
+                text_config = {}
+            elif "format" in text_config and text_config["format"] != format_config:
                raise ServiceInvalidRequestError("Conflicting response_format definitions detected.")
-            prepared_text["format"] = format_config
-            return None, prepared_text
+            text_config["format"] = format_config
+            return None, text_config

        raise ServiceInvalidRequestError("response_format must be a Pydantic model or mapping.")

@@ -245,23 +221,33 @@ class OpenAIBaseResponsesClient(OpenAIBase, BaseChatClient):

        raise ServiceInvalidRequestError("Unsupported response_format provided for Responses client.")

-    def get_conversation_id(
+    def _get_conversation_id(
        self, response: OpenAIResponse | ParsedResponse[BaseModel], store: bool | None
    ) -> str | None:
        """Get the conversation ID from the response if store is True."""
-        return None if store is False else response.id
+        if store is False:
+            return None
+        # If conversation ID exists, it means that we operate with conversation
+        # so we use conversation ID as input and output.
+        if response.conversation and response.conversation.id:
+            return response.conversation.id
+        # If conversation ID doesn't exist, we operate with responses
+        # so we use response ID as input and output.
+        return response.id

    # region Prep methods

-    def _tools_to_response_tools(
-        self, tools: Sequence[ToolProtocol | MutableMapping[str, Any]]
+    def _prepare_tools_for_openai(
+        self, tools: Sequence[ToolProtocol | MutableMapping[str, Any]] | None
    ) -> list[ToolParam | dict[str, Any]]:
        response_tools: list[ToolParam | dict[str, Any]] = []
+        if not tools:
+            return response_tools
        for tool in tools:
            if isinstance(tool, ToolProtocol):
                match tool:
                    case HostedMCPTool():
-                        response_tools.append(self.get_mcp_tool(tool))
+                        response_tools.append(self._prepare_mcp_tool(tool))
                    case HostedCodeInterpreterTool():
                        tool_args: CodeInterpreterContainerCodeInterpreterToolAuto = {"type": "auto"}
                        if tool.inputs:
@@ -363,7 +349,8 @@ class OpenAIBaseResponsesClient(OpenAIBase, BaseChatClient):
                    response_tools.append(tool_dict)
        return response_tools

-    def get_mcp_tool(self, tool: HostedMCPTool) -> Any:
+    @staticmethod
+    def _prepare_mcp_tool(tool: HostedMCPTool) -> Mcp:
        """Get MCP tool from HostedMCPTool."""
        mcp: Mcp = {
            "type": "mcp",
@@ -386,18 +373,13 @@ class OpenAIBaseResponsesClient(OpenAIBase, BaseChatClient):

        return mcp

-    async def prepare_options(
+    async def _prepare_options(
        self,
        messages: MutableSequence[ChatMessage],
        chat_options: ChatOptions,
        **kwargs: Any,
    ) -> dict[str, Any]:
        """Take ChatOptions and create the specific options for Responses API."""
-        conversation_id = kwargs.pop("conversation_id", None)
-
-        if conversation_id:
-            chat_options.conversation_id = conversation_id
-
        run_options: dict[str, Any] = chat_options.to_dict(
            exclude={
                "type",
@@ -407,12 +389,24 @@ class OpenAIBaseResponsesClient(OpenAIBase, BaseChatClient):
                "seed",  # not supported
                "stop",  # not supported
                "instructions",  # already added as system message
+                "response_format",  # handled separately
+                "conversation_id",  # handled separately
+                "additional_properties",  # handled separately
            }
        )
+        # messages
+        request_input = self._prepare_messages_for_openai(messages)
+        if not request_input:
+            raise ServiceInvalidRequestError("Messages are required for chat completions")
+        run_options["input"] = request_input

-        if chat_options.response_format:
-            run_options["response_format"] = chat_options.response_format
+        # model id
+        if not run_options.get("model"):
+            if not self.model_id:
+                raise ValueError("model_id must be a non-empty string")
+            run_options["model"] = self.model_id

+        # translations between ChatOptions and Responses API
        translations = {
            "model_id": "model",
            "allow_multiple_tool_calls": "parallel_tool_calls",
@@ -423,34 +417,53 @@ class OpenAIBaseResponsesClient(OpenAIBase, BaseChatClient):
            if old_key in run_options and old_key != new_key:
                run_options[new_key] = run_options.pop(old_key)

+        # Handle different conversation ID formats
+        if conversation_id := self._get_current_conversation_id(chat_options, **kwargs):
+            if conversation_id.startswith("resp_"):
+                # For response IDs, set previous_response_id and remove conversation property
+                run_options["previous_response_id"] = conversation_id
+            elif conversation_id.startswith("conv_"):
+                # For conversation IDs, set conversation and remove previous_response_id property
+                run_options["conversation"] = conversation_id
+            else:
+                # If the format is unrecognized, default to previous_response_id
+                run_options["previous_response_id"] = conversation_id
+
        # tools
-        if chat_options.tools is None:
-            run_options.pop("parallel_tool_calls", None)
+        if tools := self._prepare_tools_for_openai(chat_options.tools):
+            run_options["tools"] = tools
        else:
-            run_options["tools"] = self._tools_to_response_tools(chat_options.tools)
-
-        # model id
-        if not run_options.get("model"):
-            if not self.model_id:
-                raise ValueError("model_id must be a non-empty string")
-            run_options["model"] = self.model_id
-
-        # messages
-        request_input = self._prepare_chat_messages_for_request(messages)
-        if not request_input:
-            raise ServiceInvalidRequestError("Messages are required for chat completions")
-        run_options["input"] = request_input
-
-        # additional provider specific settings
-        if additional_properties := run_options.pop("additional_properties", None):
-            for key, value in additional_properties.items():
-                if value is not None:
-                    run_options[key] = value
+            run_options.pop("parallel_tool_calls", None)
+            run_options.pop("tool_choice", None)
+        # tool choice when `tool_choice` is a dict with single key `mode`, extract the mode value
        if (tool_choice := run_options.get("tool_choice")) and len(tool_choice.keys()) == 1:
            run_options["tool_choice"] = tool_choice["mode"]
+
+        # additional properties
+        additional_options = {
+            key: value for key, value in chat_options.additional_properties.items() if value is not None
+        }
+        if additional_options:
+            run_options.update(additional_options)
+
+        # response format and text config (after additional_properties so user can pass text via additional_properties)
+        response_format = chat_options.response_format
+        text_config = run_options.pop("text", None)
+        response_format, text_config = self._prepare_response_and_text_format(
+            response_format=response_format, text_config=text_config
+        )
+        if text_config:
+            run_options["text"] = text_config
+        if response_format:
+            run_options["text_format"] = response_format
+
        return run_options

-    def _prepare_chat_messages_for_request(self, chat_messages: Sequence[ChatMessage]) -> list[dict[str, Any]]:
+    def _get_current_conversation_id(self, chat_options: ChatOptions, **kwargs: Any) -> str | None:
+        """Get the current conversation ID from chat options or kwargs."""
+        return chat_options.conversation_id or kwargs.get("conversation_id")
+
+    def _prepare_messages_for_openai(self, chat_messages: Sequence[ChatMessage]) -> list[dict[str, Any]]:
        """Prepare the chat messages for a request.

        Allowing customization of the key names for role/author, and optionally overriding the role.
@@ -476,16 +489,16 @@ class OpenAIBaseResponsesClient(OpenAIBase, BaseChatClient):
                    and "fc_id" in content.additional_properties
                ):
                    call_id_to_id[content.call_id] = content.additional_properties["fc_id"]
-        list_of_list = [self._openai_chat_message_parser(message, call_id_to_id) for message in chat_messages]
+        list_of_list = [self._prepare_message_for_openai(message, call_id_to_id) for message in chat_messages]
        # Flatten the list of lists into a single list
        return list(chain.from_iterable(list_of_list))

-    def _openai_chat_message_parser(
+    def _prepare_message_for_openai(
        self,
        message: ChatMessage,
        call_id_to_id: dict[str, str],
    ) -> list[dict[str, Any]]:
-        """Parse a chat message into the openai format."""
+        """Prepare a chat message for the OpenAI Responses API format."""
        all_messages: list[dict[str, Any]] = []
        args: dict[str, Any] = {
            "role": message.role.value if isinstance(message.role, Role) else message.role,
@@ -497,28 +510,28 @@ class OpenAIBaseResponsesClient(OpenAIBase, BaseChatClient):
                    continue
                case FunctionResultContent():
                    new_args: dict[str, Any] = {}
-                    new_args.update(self._openai_content_parser(message.role, content, call_id_to_id))
+                    new_args.update(self._prepare_content_for_openai(message.role, content, call_id_to_id))
                    all_messages.append(new_args)
                case FunctionCallContent():
-                    function_call = self._openai_content_parser(message.role, content, call_id_to_id)
+                    function_call = self._prepare_content_for_openai(message.role, content, call_id_to_id)
                    all_messages.append(function_call)  # type: ignore
                case FunctionApprovalResponseContent() | FunctionApprovalRequestContent():
-                    all_messages.append(self._openai_content_parser(message.role, content, call_id_to_id))  # type: ignore
+                    all_messages.append(self._prepare_content_for_openai(message.role, content, call_id_to_id))  # type: ignore
                case _:
                    if "content" not in args:
                        args["content"] = []
-                    args["content"].append(self._openai_content_parser(message.role, content, call_id_to_id))  # type: ignore
+                    args["content"].append(self._prepare_content_for_openai(message.role, content, call_id_to_id))  # type: ignore
        if "content" in args or "tool_calls" in args:
            all_messages.append(args)
        return all_messages

-    def _openai_content_parser(
+    def _prepare_content_for_openai(
        self,
        role: Role,
        content: Contents,
        call_id_to_id: dict[str, str],
    ) -> dict[str, Any]:
-        """Parse contents into the openai format."""
+        """Prepare content for the OpenAI Responses API format."""
        match content:
            case TextContent():
                return {
@@ -625,14 +638,13 @@ class OpenAIBaseResponsesClient(OpenAIBase, BaseChatClient):
                logger.debug("Unsupported content type passed (type: %s)", type(content))
                return {}

-    # region Response creation methods
-
-    def _create_response_content(
+    # region Parse methods
+    def _parse_response_from_openai(
        self,
        response: OpenAIResponse | ParsedResponse[BaseModel],
        chat_options: ChatOptions,
    ) -> "ChatResponse":
-        """Create a chat message content object from a choice."""
+        """Parse an OpenAI Responses API response into a ChatResponse."""
        structured_response: BaseModel | None = response.output_parsed if isinstance(response, ParsedResponse) else None  # type: ignore[reportUnknownMemberType]

        metadata: dict[str, Any] = response.metadata or {}
@@ -826,11 +838,9 @@ class OpenAIBaseResponsesClient(OpenAIBase, BaseChatClient):
            "raw_representation": response,
        }

-        conversation_id = self.get_conversation_id(response, chat_options.store)  # type: ignore[reportArgumentType]
-
-        if conversation_id:
+        if conversation_id := self._get_conversation_id(response, chat_options.store):
            args["conversation_id"] = conversation_id
-        if response.usage and (usage_details := self._usage_details_from_openai(response.usage)):
+        if response.usage and (usage_details := self._parse_usage_from_openai(response.usage)):
            args["usage_details"] = usage_details
        if structured_response:
            args["value"] = structured_response
@@ -838,16 +848,17 @@ class OpenAIBaseResponsesClient(OpenAIBase, BaseChatClient):
            args["response_format"] = chat_options.response_format
        return ChatResponse(**args)

-    def _create_streaming_response_content(
+    def _parse_chunk_from_openai(
        self,
        event: OpenAIResponseStreamEvent,
        chat_options: ChatOptions,
        function_call_ids: dict[int, tuple[str, str]],
    ) -> ChatResponseUpdate:
-        """Create a streaming chat message content object from a choice."""
+        """Parse an OpenAI Responses API streaming event into a ChatResponseUpdate."""
        metadata: dict[str, Any] = {}
        contents: list[Contents] = []
        conversation_id: str | None = None
+        response_id: str | None = None
        model = self.model_id
        # TODO(peterychang): Add support for other content types
        match event.type:
@@ -930,11 +941,18 @@ class OpenAIBaseResponsesClient(OpenAIBase, BaseChatClient):
            case "response.reasoning_summary_text.done":
                contents.append(TextReasoningContent(text=event.text, raw_representation=event))
                metadata.update(self._get_metadata_from_response(event))
+            case "response.created":
+                response_id = event.response.id
+                conversation_id = self._get_conversation_id(event.response, chat_options.store)
+            case "response.in_progress":
+                response_id = event.response.id
+                conversation_id = self._get_conversation_id(event.response, chat_options.store)
            case "response.completed":
-                conversation_id = self.get_conversation_id(event.response, chat_options.store)
+                response_id = event.response.id
+                conversation_id = self._get_conversation_id(event.response, chat_options.store)
                model = event.response.model
                if event.response.usage:
-                    usage = self._usage_details_from_openai(event.response.usage)
+                    usage = self._parse_usage_from_openai(event.response.usage)
                    if usage:
                        contents.append(UsageContent(details=usage, raw_representation=event))
            case "response.output_item.added":
@@ -1096,13 +1114,14 @@ class OpenAIBaseResponsesClient(OpenAIBase, BaseChatClient):
        return ChatResponseUpdate(
            contents=contents,
            conversation_id=conversation_id,
+            response_id=response_id,
            role=Role.ASSISTANT,
            model_id=model,
            additional_properties=metadata,
            raw_representation=event,
        )

-    def _usage_details_from_openai(self, usage: ResponseUsage) -> UsageDetails | None:
+    def _parse_usage_from_openai(self, usage: ResponseUsage) -> UsageDetails | None:
        details = UsageDetails(
            input_token_count=usage.input_tokens,
            output_token_count=usage.output_tokens,
@@ -160,16 +160,16 @@ class OpenAIBase(SerializationMixin):
        for key, value in kwargs.items():
            setattr(self, key, value)

-    async def initialize_client(self) -> None:
+    async def _initialize_client(self) -> None:
        """Initialize OpenAI client asynchronously.

        Override in subclasses to initialize the OpenAI client asynchronously.
        """
        pass

-    async def ensure_client(self) -> AsyncOpenAI:
+    async def _ensure_client(self) -> AsyncOpenAI:
        """Ensure OpenAI client is initialized."""
-        await self.initialize_client()
+        await self._initialize_client()
        if self.client is None:
            raise ServiceInitializationError("OpenAI client is not initialized")

@@ -4,7 +4,7 @@ description = "Microsoft Agent Framework for building AI Agents with Python. Thi
 authors = [{ name = "Microsoft", email = "af-support@microsoft.com"}]
 readme = "README.md"
 requires-python = ">=3.10"
-version = "1.0.0b251216"
+version = "1.0.0b251223"
 license-files = ["LICENSE"]
 urls.homepage = "https://aka.ms/agent-framework"
 urls.source = "https://github.com/microsoft/agent-framework/tree/main/python"
@@ -53,6 +53,7 @@ all = [
    "agent-framework-durabletask",
    "agent-framework-lab",
    "agent-framework-mem0",
+    "agent-framework-ollama",
    "agent-framework-purview",
    "agent-framework-redis",
 ]
@@ -193,7 +193,7 @@ async def test_cmc(
    mock_create.assert_awaited_once_with(
        model=azure_openai_unit_test_env["AZURE_OPENAI_CHAT_DEPLOYMENT_NAME"],
        stream=False,
-        messages=azure_chat_client._prepare_chat_history_for_request(chat_history),  # type: ignore
+        messages=azure_chat_client._prepare_messages_for_openai(chat_history),  # type: ignore
    )


@@ -216,7 +216,7 @@ async def test_cmc_with_logit_bias(

    mock_create.assert_awaited_once_with(
        model=azure_openai_unit_test_env["AZURE_OPENAI_CHAT_DEPLOYMENT_NAME"],
-        messages=azure_chat_client._prepare_chat_history_for_request(chat_history),  # type: ignore
+        messages=azure_chat_client._prepare_messages_for_openai(chat_history),  # type: ignore
        stream=False,
        logit_bias=token_bias,
    )
@@ -241,7 +241,7 @@ async def test_cmc_with_stop(

    mock_create.assert_awaited_once_with(
        model=azure_openai_unit_test_env["AZURE_OPENAI_CHAT_DEPLOYMENT_NAME"],
-        messages=azure_chat_client._prepare_chat_history_for_request(chat_history),  # type: ignore
+        messages=azure_chat_client._prepare_messages_for_openai(chat_history),  # type: ignore
        stream=False,
        stop=stop,
    )
@@ -311,7 +311,7 @@ async def test_azure_on_your_data(

    mock_create.assert_awaited_once_with(
        model=azure_openai_unit_test_env["AZURE_OPENAI_CHAT_DEPLOYMENT_NAME"],
-        messages=azure_chat_client._prepare_chat_history_for_request(messages_out),  # type: ignore
+        messages=azure_chat_client._prepare_messages_for_openai(messages_out),  # type: ignore
        stream=False,
        extra_body=expected_data_settings,
    )
@@ -381,7 +381,7 @@ async def test_azure_on_your_data_string(

    mock_create.assert_awaited_once_with(
        model=azure_openai_unit_test_env["AZURE_OPENAI_CHAT_DEPLOYMENT_NAME"],
-        messages=azure_chat_client._prepare_chat_history_for_request(messages_out),  # type: ignore
+        messages=azure_chat_client._prepare_messages_for_openai(messages_out),  # type: ignore
        stream=False,
        extra_body=expected_data_settings,
    )
@@ -438,7 +438,7 @@ async def test_azure_on_your_data_fail(

    mock_create.assert_awaited_once_with(
        model=azure_openai_unit_test_env["AZURE_OPENAI_CHAT_DEPLOYMENT_NAME"],
-        messages=azure_chat_client._prepare_chat_history_for_request(messages_out),  # type: ignore
+        messages=azure_chat_client._prepare_messages_for_openai(messages_out),  # type: ignore
        stream=False,
        extra_body=expected_data_settings,
    )
@@ -584,7 +584,7 @@ async def test_get_streaming(
    mock_create.assert_awaited_once_with(
        model=azure_openai_unit_test_env["AZURE_OPENAI_CHAT_DEPLOYMENT_NAME"],
        stream=True,
-        messages=azure_chat_client._prepare_chat_history_for_request(chat_history),  # type: ignore
+        messages=azure_chat_client._prepare_messages_for_openai(chat_history),  # type: ignore
        # NOTE: The `stream_options={"include_usage": True}` is explicitly enforced in
        # `OpenAIChatCompletionBase._inner_get_streaming_response`.
        # To ensure consistency, we align the arguments here accordingly.
@@ -24,14 +24,14 @@ from agent_framework import (
 )
 from agent_framework._mcp import (
    MCPTool,
-    _ai_content_to_mcp_types,
-    _chat_message_to_mcp_types,
    _get_input_model_from_mcp_prompt,
    _get_input_model_from_mcp_tool,
-    _mcp_call_tool_result_to_ai_contents,
-    _mcp_prompt_message_to_chat_message,
-    _mcp_type_to_ai_content,
    _normalize_mcp_name,
+    _parse_content_from_mcp,
+    _parse_contents_from_mcp_tool_result,
+    _parse_message_from_mcp,
+    _prepare_content_for_mcp,
+    _prepare_message_for_mcp,
 )
 from agent_framework.exceptions import ToolException, ToolExecutionException

@@ -60,7 +60,7 @@ def test_normalize_mcp_name():
 def test_mcp_prompt_message_to_ai_content():
    """Test conversion from MCP prompt message to AI content."""
    mcp_message = types.PromptMessage(role="user", content=types.TextContent(type="text", text="Hello, world!"))
-    ai_content = _mcp_prompt_message_to_chat_message(mcp_message)
+    ai_content = _parse_message_from_mcp(mcp_message)

    assert isinstance(ai_content, ChatMessage)
    assert ai_content.role.value == "user"
@@ -70,7 +70,7 @@ def test_mcp_prompt_message_to_ai_content():
    assert ai_content.raw_representation == mcp_message


-def test_mcp_call_tool_result_to_ai_contents():
+def test_parse_contents_from_mcp_tool_result():
    """Test conversion from MCP tool result to AI contents."""
    mcp_result = types.CallToolResult(
        content=[
@@ -79,7 +79,7 @@ def test_mcp_call_tool_result_to_ai_contents():
            types.ImageContent(type="image", data=b"abc", mimeType="image/webp"),
        ]
    )
-    ai_contents = _mcp_call_tool_result_to_ai_contents(mcp_result)
+    ai_contents = _parse_contents_from_mcp_tool_result(mcp_result)

    assert len(ai_contents) == 3
    assert isinstance(ai_contents[0], TextContent)
@@ -100,7 +100,7 @@ def test_mcp_call_tool_result_with_meta_error():
        _meta={"isError": True, "errorCode": "TOOL_ERROR", "errorMessage": "Tool execution failed"},
    )

-    ai_contents = _mcp_call_tool_result_to_ai_contents(mcp_result)
+    ai_contents = _parse_contents_from_mcp_tool_result(mcp_result)

    assert len(ai_contents) == 1
    assert isinstance(ai_contents[0], TextContent)
@@ -131,7 +131,7 @@ def test_mcp_call_tool_result_with_meta_arbitrary_data():
        },
    )

-    ai_contents = _mcp_call_tool_result_to_ai_contents(mcp_result)
+    ai_contents = _parse_contents_from_mcp_tool_result(mcp_result)

    assert len(ai_contents) == 1
    assert isinstance(ai_contents[0], TextContent)
@@ -153,7 +153,7 @@ def test_mcp_call_tool_result_with_meta_merging_existing_properties():
    text_content = types.TextContent(type="text", text="Test content")
    mcp_result = types.CallToolResult(content=[text_content], _meta={"newField": "newValue", "isError": False})

-    ai_contents = _mcp_call_tool_result_to_ai_contents(mcp_result)
+    ai_contents = _parse_contents_from_mcp_tool_result(mcp_result)

    assert len(ai_contents) == 1
    content = ai_contents[0]
@@ -169,7 +169,7 @@ def test_mcp_call_tool_result_with_meta_none():
    mcp_result = types.CallToolResult(content=[types.TextContent(type="text", text="No meta test")])
    # No _meta field set

-    ai_contents = _mcp_call_tool_result_to_ai_contents(mcp_result)
+    ai_contents = _parse_contents_from_mcp_tool_result(mcp_result)

    assert len(ai_contents) == 1
    assert isinstance(ai_contents[0], TextContent)
@@ -191,7 +191,7 @@ def test_mcp_call_tool_result_regression_successful_workflow():
        ]
    )

-    ai_contents = _mcp_call_tool_result_to_ai_contents(mcp_result)
+    ai_contents = _parse_contents_from_mcp_tool_result(mcp_result)

    # Verify basic conversion still works correctly
    assert len(ai_contents) == 2
@@ -213,7 +213,7 @@ def test_mcp_call_tool_result_regression_successful_workflow():
 def test_mcp_content_types_to_ai_content_text():
    """Test conversion of MCP text content to AI content."""
    mcp_content = types.TextContent(type="text", text="Sample text")
-    ai_content = _mcp_type_to_ai_content(mcp_content)[0]
+    ai_content = _parse_content_from_mcp(mcp_content)[0]

    assert isinstance(ai_content, TextContent)
    assert ai_content.text == "Sample text"
@@ -224,7 +224,7 @@ def test_mcp_content_types_to_ai_content_image():
    """Test conversion of MCP image content to AI content."""
    mcp_content = types.ImageContent(type="image", data="abc", mimeType="image/jpeg")
    mcp_content = types.ImageContent(type="image", data=b"abc", mimeType="image/jpeg")
-    ai_content = _mcp_type_to_ai_content(mcp_content)[0]
+    ai_content = _parse_content_from_mcp(mcp_content)[0]

    assert isinstance(ai_content, DataContent)
    assert ai_content.uri == "data:image/jpeg;base64,abc"
@@ -235,7 +235,7 @@ def test_mcp_content_types_to_ai_content_image():
 def test_mcp_content_types_to_ai_content_audio():
    """Test conversion of MCP audio content to AI content."""
    mcp_content = types.AudioContent(type="audio", data="def", mimeType="audio/wav")
-    ai_content = _mcp_type_to_ai_content(mcp_content)[0]
+    ai_content = _parse_content_from_mcp(mcp_content)[0]

    assert isinstance(ai_content, DataContent)
    assert ai_content.uri == "data:audio/wav;base64,def"
@@ -251,7 +251,7 @@ def test_mcp_content_types_to_ai_content_resource_link():
        name="test_resource",
        mimeType="application/json",
    )
-    ai_content = _mcp_type_to_ai_content(mcp_content)[0]
+    ai_content = _parse_content_from_mcp(mcp_content)[0]

    assert isinstance(ai_content, UriContent)
    assert ai_content.uri == "https://example.com/resource"
@@ -267,7 +267,7 @@ def test_mcp_content_types_to_ai_content_embedded_resource_text():
        text="Embedded text content",
    )
    mcp_content = types.EmbeddedResource(type="resource", resource=text_resource)
-    ai_content = _mcp_type_to_ai_content(mcp_content)[0]
+    ai_content = _parse_content_from_mcp(mcp_content)[0]

    assert isinstance(ai_content, TextContent)
    assert ai_content.text == "Embedded text content"
@@ -283,7 +283,7 @@ def test_mcp_content_types_to_ai_content_embedded_resource_blob():
        blob="data:application/octet-stream;base64,dGVzdCBkYXRh",
    )
    mcp_content = types.EmbeddedResource(type="resource", resource=blob_resource)
-    ai_content = _mcp_type_to_ai_content(mcp_content)[0]
+    ai_content = _parse_content_from_mcp(mcp_content)[0]

    assert isinstance(ai_content, DataContent)
    assert ai_content.uri == "data:application/octet-stream;base64,dGVzdCBkYXRh"
@@ -294,7 +294,7 @@ def test_mcp_content_types_to_ai_content_embedded_resource_blob():
 def test_ai_content_to_mcp_content_types_text():
    """Test conversion of AI text content to MCP content."""
    ai_content = TextContent(text="Sample text")
-    mcp_content = _ai_content_to_mcp_types(ai_content)
+    mcp_content = _prepare_content_for_mcp(ai_content)

    assert isinstance(mcp_content, types.TextContent)
    assert mcp_content.type == "text"
@@ -304,7 +304,7 @@ def test_ai_content_to_mcp_content_types_text():
 def test_ai_content_to_mcp_content_types_data_image():
    """Test conversion of AI data content to MCP content."""
    ai_content = DataContent(uri="data:image/png;base64,xyz", media_type="image/png")
-    mcp_content = _ai_content_to_mcp_types(ai_content)
+    mcp_content = _prepare_content_for_mcp(ai_content)

    assert isinstance(mcp_content, types.ImageContent)
    assert mcp_content.type == "image"
@@ -315,7 +315,7 @@ def test_ai_content_to_mcp_content_types_data_image():
 def test_ai_content_to_mcp_content_types_data_audio():
    """Test conversion of AI data content to MCP content."""
    ai_content = DataContent(uri="data:audio/mpeg;base64,xyz", media_type="audio/mpeg")
-    mcp_content = _ai_content_to_mcp_types(ai_content)
+    mcp_content = _prepare_content_for_mcp(ai_content)

    assert isinstance(mcp_content, types.AudioContent)
    assert mcp_content.type == "audio"
@@ -329,7 +329,7 @@ def test_ai_content_to_mcp_content_types_data_binary():
        uri="data:application/octet-stream;base64,xyz",
        media_type="application/octet-stream",
    )
-    mcp_content = _ai_content_to_mcp_types(ai_content)
+    mcp_content = _prepare_content_for_mcp(ai_content)

    assert isinstance(mcp_content, types.EmbeddedResource)
    assert mcp_content.type == "resource"
@@ -340,7 +340,7 @@ def test_ai_content_to_mcp_content_types_data_binary():
 def test_ai_content_to_mcp_content_types_uri():
    """Test conversion of AI URI content to MCP content."""
    ai_content = UriContent(uri="https://example.com/resource", media_type="application/json")
-    mcp_content = _ai_content_to_mcp_types(ai_content)
+    mcp_content = _prepare_content_for_mcp(ai_content)

    assert isinstance(mcp_content, types.ResourceLink)
    assert mcp_content.type == "resource_link"
@@ -348,7 +348,7 @@ def test_ai_content_to_mcp_content_types_uri():
    assert mcp_content.mimeType == "application/json"


-def test_chat_message_to_mcp_types():
+def test_prepare_message_for_mcp():
    message = ChatMessage(
        role="user",
        contents=[
@@ -356,7 +356,7 @@ def test_chat_message_to_mcp_types():
            DataContent(uri="data:image/png;base64,xyz", media_type="image/png"),
        ],
    )
-    mcp_contents = _chat_message_to_mcp_types(message)
+    mcp_contents = _prepare_message_for_mcp(message)
    assert len(mcp_contents) == 2
    assert isinstance(mcp_contents[0], types.TextContent)
    assert isinstance(mcp_contents[1], types.ImageContent)
@@ -463,9 +463,9 @@ async def test_openai_assistants_client_process_stream_events_requires_action(mo
    """Test _process_stream_events with thread.run.requires_action event."""
    chat_client = create_test_openai_assistants_client(mock_async_openai)

-    # Mock the _create_function_call_contents method to return test content
+    # Mock the _parse_function_calls_from_assistants method to return test content
    test_function_content = FunctionCallContent(call_id="call-123", name="test_func", arguments={"arg": "value"})
-    chat_client._create_function_call_contents = MagicMock(return_value=[test_function_content])  # type: ignore
+    chat_client._parse_function_calls_from_assistants = MagicMock(return_value=[test_function_content])  # type: ignore

    # Create a mock Run object
    mock_run = MagicMock(spec=Run)
@@ -498,8 +498,8 @@ async def test_openai_assistants_client_process_stream_events_requires_action(mo
    assert update.contents[0] == test_function_content
    assert update.raw_representation == mock_run

-    # Verify _create_function_call_contents was called correctly
-    chat_client._create_function_call_contents.assert_called_once_with(mock_run, None)  # type: ignore
+    # Verify _parse_function_calls_from_assistants was called correctly
+    chat_client._parse_function_calls_from_assistants.assert_called_once_with(mock_run, None)  # type: ignore


 async def test_openai_assistants_client_process_stream_events_run_step_created(mock_async_openai: MagicMock) -> None:
@@ -585,8 +585,8 @@ async def test_openai_assistants_client_process_stream_events_run_completed_with
    assert update.raw_representation == mock_run


-def test_openai_assistants_client_create_function_call_contents_basic(mock_async_openai: MagicMock) -> None:
-    """Test _create_function_call_contents with a simple function call."""
+def test_openai_assistants_client_parse_function_calls_from_assistants_basic(mock_async_openai: MagicMock) -> None:
+    """Test _parse_function_calls_from_assistants with a simple function call."""

    chat_client = create_test_openai_assistants_client(mock_async_openai)

@@ -605,7 +605,7 @@ def test_openai_assistants_client_create_function_call_contents_basic(mock_async

    # Call the method
    response_id = "response_456"
-    contents = chat_client._create_function_call_contents(mock_run, response_id)  # type: ignore
+    contents = chat_client._parse_function_calls_from_assistants(mock_run, response_id)  # type: ignore

    # Test that one function call content was created
    assert len(contents) == 1
@@ -825,24 +825,24 @@ def test_openai_assistants_client_prepare_options_with_image_content(mock_async_
    assert message["content"][0]["image_url"]["url"] == "https://example.com/image.jpg"


-def test_openai_assistants_client_convert_function_results_to_tool_output_empty(mock_async_openai: MagicMock) -> None:
-    """Test _convert_function_results_to_tool_output with empty list."""
+def test_openai_assistants_client_prepare_tool_outputs_for_assistants_empty(mock_async_openai: MagicMock) -> None:
+    """Test _prepare_tool_outputs_for_assistants with empty list."""
    chat_client = create_test_openai_assistants_client(mock_async_openai)

-    run_id, tool_outputs = chat_client._convert_function_results_to_tool_output([])  # type: ignore
+    run_id, tool_outputs = chat_client._prepare_tool_outputs_for_assistants([])  # type: ignore

    assert run_id is None
    assert tool_outputs is None


-def test_openai_assistants_client_convert_function_results_to_tool_output_valid(mock_async_openai: MagicMock) -> None:
-    """Test _convert_function_results_to_tool_output with valid function results."""
+def test_openai_assistants_client_prepare_tool_outputs_for_assistants_valid(mock_async_openai: MagicMock) -> None:
+    """Test _prepare_tool_outputs_for_assistants with valid function results."""
    chat_client = create_test_openai_assistants_client(mock_async_openai)

    call_id = json.dumps(["run-123", "call-456"])
    function_result = FunctionResultContent(call_id=call_id, result="Function executed successfully")

-    run_id, tool_outputs = chat_client._convert_function_results_to_tool_output([function_result])  # type: ignore
+    run_id, tool_outputs = chat_client._prepare_tool_outputs_for_assistants([function_result])  # type: ignore

    assert run_id == "run-123"
    assert tool_outputs is not None
@@ -851,10 +851,10 @@ def test_openai_assistants_client_convert_function_results_to_tool_output_valid(
    assert tool_outputs[0].get("output") == "Function executed successfully"


-def test_openai_assistants_client_convert_function_results_to_tool_output_mismatched_run_ids(
+def test_openai_assistants_client_prepare_tool_outputs_for_assistants_mismatched_run_ids(
    mock_async_openai: MagicMock,
 ) -> None:
-    """Test _convert_function_results_to_tool_output with mismatched run IDs."""
+    """Test _prepare_tool_outputs_for_assistants with mismatched run IDs."""
    chat_client = create_test_openai_assistants_client(mock_async_openai)

    # Create function results with different run IDs
@@ -863,7 +863,7 @@ def test_openai_assistants_client_convert_function_results_to_tool_output_mismat
    function_result1 = FunctionResultContent(call_id=call_id1, result="Result 1")
    function_result2 = FunctionResultContent(call_id=call_id2, result="Result 2")

-    run_id, tool_outputs = chat_client._convert_function_results_to_tool_output([function_result1, function_result2])  # type: ignore
+    run_id, tool_outputs = chat_client._prepare_tool_outputs_for_assistants([function_result1, function_result2])  # type: ignore

    # Should only process the first one since run IDs don't match
    assert run_id == "run-123"
@@ -182,12 +182,12 @@ def test_unsupported_tool_handling(openai_unit_test_env: dict[str, str]) -> None
    unsupported_tool.__class__.__name__ = "UnsupportedAITool"

    # This should ignore the unsupported ToolProtocol and return empty list
-    result = client._chat_to_tool_spec([unsupported_tool])  # type: ignore
+    result = client._prepare_tools_for_openai([unsupported_tool])  # type: ignore
    assert result == []

    # Also test with a non-ToolProtocol that should be converted to dict
    dict_tool = {"type": "function", "name": "test"}
-    result = client._chat_to_tool_spec([dict_tool])  # type: ignore
+    result = client._prepare_tools_for_openai([dict_tool])  # type: ignore
    assert result == [dict_tool]


@@ -637,7 +637,7 @@ def test_chat_response_content_order_text_before_tool_calls(openai_unit_test_env
    )

    client = OpenAIChatClient()
-    response = client._create_chat_response(mock_response, ChatOptions())
+    response = client._parse_response_from_openai(mock_response, ChatOptions())

    # Verify we have both text and tool call content
    assert len(response.messages) == 1
@@ -658,7 +658,7 @@ def test_function_result_falsy_values_handling(openai_unit_test_env: dict[str, s
    # Test with empty list (falsy but not None)
    message_with_empty_list = ChatMessage(role="tool", contents=[FunctionResultContent(call_id="call-123", result=[])])

-    openai_messages = client._openai_chat_message_parser(message_with_empty_list)
+    openai_messages = client._prepare_message_for_openai(message_with_empty_list)
    assert len(openai_messages) == 1
    assert openai_messages[0]["content"] == "[]"  # Empty list should be JSON serialized

@@ -667,14 +667,14 @@ def test_function_result_falsy_values_handling(openai_unit_test_env: dict[str, s
        role="tool", contents=[FunctionResultContent(call_id="call-456", result="")]
    )

-    openai_messages = client._openai_chat_message_parser(message_with_empty_string)
+    openai_messages = client._prepare_message_for_openai(message_with_empty_string)
    assert len(openai_messages) == 1
    assert openai_messages[0]["content"] == ""  # Empty string should be preserved

    # Test with False (falsy but not None)
    message_with_false = ChatMessage(role="tool", contents=[FunctionResultContent(call_id="call-789", result=False)])

-    openai_messages = client._openai_chat_message_parser(message_with_false)
+    openai_messages = client._prepare_message_for_openai(message_with_false)
    assert len(openai_messages) == 1
    assert openai_messages[0]["content"] == "false"  # False should be JSON serialized

@@ -695,7 +695,7 @@ def test_function_result_exception_handling(openai_unit_test_env: dict[str, str]
        ],
    )

-    openai_messages = client._openai_chat_message_parser(message_with_exception)
+    openai_messages = client._prepare_message_for_openai(message_with_exception)
    assert len(openai_messages) == 1
    assert openai_messages[0]["content"] == "Error: Function failed."
    assert openai_messages[0]["tool_call_id"] == "call-123"
@@ -708,8 +708,8 @@ def test_prepare_function_call_results_string_passthrough():
    assert isinstance(result, str)


-def test_openai_content_parser_data_content_image(openai_unit_test_env: dict[str, str]) -> None:
-    """Test _openai_content_parser converts DataContent with image media type to OpenAI format."""
+def test_prepare_content_for_openai_data_content_image(openai_unit_test_env: dict[str, str]) -> None:
+    """Test _prepare_content_for_openai converts DataContent with image media type to OpenAI format."""
    client = OpenAIChatClient()

    # Test DataContent with image media type
@@ -718,7 +718,7 @@ def test_openai_content_parser_data_content_image(openai_unit_test_env: dict[str
        media_type="image/png",
    )

-    result = client._openai_content_parser(image_data_content)  # type: ignore
+    result = client._prepare_content_for_openai(image_data_content)  # type: ignore

    # Should convert to OpenAI image_url format
    assert result["type"] == "image_url"
@@ -727,7 +727,7 @@ def test_openai_content_parser_data_content_image(openai_unit_test_env: dict[str
    # Test DataContent with non-image media type should use default model_dump
    text_data_content = DataContent(uri="data:text/plain;base64,SGVsbG8gV29ybGQ=", media_type="text/plain")

-    result = client._openai_content_parser(text_data_content)  # type: ignore
+    result = client._prepare_content_for_openai(text_data_content)  # type: ignore

    # Should use default model_dump format
    assert result["type"] == "data"
@@ -740,7 +740,7 @@ def test_openai_content_parser_data_content_image(openai_unit_test_env: dict[str
        media_type="audio/wav",
    )

-    result = client._openai_content_parser(audio_data_content)  # type: ignore
+    result = client._prepare_content_for_openai(audio_data_content)  # type: ignore

    # Should convert to OpenAI input_audio format
    assert result["type"] == "input_audio"
@@ -751,7 +751,7 @@ def test_openai_content_parser_data_content_image(openai_unit_test_env: dict[str
    # Test DataContent with MP3 audio
    mp3_data_content = DataContent(uri="data:audio/mp3;base64,//uQAAAAWGluZwAAAA8AAAACAAACcQ==", media_type="audio/mp3")

-    result = client._openai_content_parser(mp3_data_content)  # type: ignore
+    result = client._prepare_content_for_openai(mp3_data_content)  # type: ignore

    # Should convert to OpenAI input_audio format with mp3
    assert result["type"] == "input_audio"
@@ -760,8 +760,8 @@ def test_openai_content_parser_data_content_image(openai_unit_test_env: dict[str
    assert result["input_audio"]["format"] == "mp3"


-def test_openai_content_parser_document_file_mapping(openai_unit_test_env: dict[str, str]) -> None:
-    """Test _openai_content_parser converts document files (PDF, DOCX, etc.) to OpenAI file format."""
+def test_prepare_content_for_openai_document_file_mapping(openai_unit_test_env: dict[str, str]) -> None:
+    """Test _prepare_content_for_openai converts document files (PDF, DOCX, etc.) to OpenAI file format."""
    client = OpenAIChatClient()

    # Test PDF without filename - should omit filename in OpenAI payload
@@ -770,7 +770,7 @@ def test_openai_content_parser_document_file_mapping(openai_unit_test_env: dict[
        media_type="application/pdf",
    )

-    result = client._openai_content_parser(pdf_data_content)  # type: ignore
+    result = client._prepare_content_for_openai(pdf_data_content)  # type: ignore

    # Should convert to OpenAI file format without filename
    assert result["type"] == "file"
@@ -787,7 +787,7 @@ def test_openai_content_parser_document_file_mapping(openai_unit_test_env: dict[
        additional_properties={"filename": "report.pdf"},
    )

-    result = client._openai_content_parser(pdf_with_filename)  # type: ignore
+    result = client._prepare_content_for_openai(pdf_with_filename)  # type: ignore

    # Should use custom filename
    assert result["type"] == "file"
@@ -820,7 +820,7 @@ def test_openai_content_parser_document_file_mapping(openai_unit_test_env: dict[
            media_type=case["media_type"],
        )

-        result = client._openai_content_parser(doc_content)  # type: ignore
+        result = client._prepare_content_for_openai(doc_content)  # type: ignore

        # All application/* types should now be mapped to file format
        assert result["type"] == "file"
@@ -834,7 +834,7 @@ def test_openai_content_parser_document_file_mapping(openai_unit_test_env: dict[
            additional_properties={"filename": case["filename"]},
        )

-        result = client._openai_content_parser(doc_with_filename)  # type: ignore
+        result = client._prepare_content_for_openai(doc_with_filename)  # type: ignore

        # Should now use file format with filename
        assert result["type"] == "file"
@@ -848,7 +848,7 @@ def test_openai_content_parser_document_file_mapping(openai_unit_test_env: dict[
        additional_properties={},
    )

-    result = client._openai_content_parser(pdf_empty_props)  # type: ignore
+    result = client._prepare_content_for_openai(pdf_empty_props)  # type: ignore

    assert result["type"] == "file"
    assert "filename" not in result["file"]
@@ -860,7 +860,7 @@ def test_openai_content_parser_document_file_mapping(openai_unit_test_env: dict[
        additional_properties={"filename": None},
    )

-    result = client._openai_content_parser(pdf_none_filename)  # type: ignore
+    result = client._prepare_content_for_openai(pdf_none_filename)  # type: ignore

    assert result["type"] == "file"
    assert "filename" not in result["file"]  # None filename should be omitted
@@ -76,7 +76,7 @@ async def test_cmc(
    mock_create.assert_awaited_once_with(
        model=openai_unit_test_env["OPENAI_CHAT_MODEL_ID"],
        stream=False,
-        messages=openai_chat_completion._prepare_chat_history_for_request(chat_history),  # type: ignore
+        messages=openai_chat_completion._prepare_messages_for_openai(chat_history),  # type: ignore
    )


@@ -97,7 +97,7 @@ async def test_cmc_chat_options(
    mock_create.assert_awaited_once_with(
        model=openai_unit_test_env["OPENAI_CHAT_MODEL_ID"],
        stream=False,
-        messages=openai_chat_completion._prepare_chat_history_for_request(chat_history),  # type: ignore
+        messages=openai_chat_completion._prepare_messages_for_openai(chat_history),  # type: ignore
    )


@@ -120,7 +120,7 @@ async def test_cmc_no_fcc_in_response(
    mock_create.assert_awaited_once_with(
        model=openai_unit_test_env["OPENAI_CHAT_MODEL_ID"],
        stream=False,
-        messages=openai_chat_completion._prepare_chat_history_for_request(orig_chat_history),  # type: ignore
+        messages=openai_chat_completion._prepare_messages_for_openai(orig_chat_history),  # type: ignore
    )


@@ -167,7 +167,7 @@ async def test_scmc_chat_options(
        model=openai_unit_test_env["OPENAI_CHAT_MODEL_ID"],
        stream=True,
        stream_options={"include_usage": True},
-        messages=openai_chat_completion._prepare_chat_history_for_request(chat_history),  # type: ignore
+        messages=openai_chat_completion._prepare_messages_for_openai(chat_history),  # type: ignore
    )


@@ -203,7 +203,7 @@ async def test_cmc_additional_properties(
    mock_create.assert_awaited_once_with(
        model=openai_unit_test_env["OPENAI_CHAT_MODEL_ID"],
        stream=False,
-        messages=openai_chat_completion._prepare_chat_history_for_request(chat_history),  # type: ignore
+        messages=openai_chat_completion._prepare_messages_for_openai(chat_history),  # type: ignore
        reasoning_effort="low",
    )

@@ -246,7 +246,7 @@ async def test_get_streaming(
        model=openai_unit_test_env["OPENAI_CHAT_MODEL_ID"],
        stream=True,
        stream_options={"include_usage": True},
-        messages=openai_chat_completion._prepare_chat_history_for_request(orig_chat_history),  # type: ignore
+        messages=openai_chat_completion._prepare_messages_for_openai(orig_chat_history),  # type: ignore
    )


@@ -285,7 +285,7 @@ async def test_get_streaming_singular(
        model=openai_unit_test_env["OPENAI_CHAT_MODEL_ID"],
        stream=True,
        stream_options={"include_usage": True},
-        messages=openai_chat_completion._prepare_chat_history_for_request(orig_chat_history),  # type: ignore
+        messages=openai_chat_completion._prepare_messages_for_openai(orig_chat_history),  # type: ignore
    )


@@ -349,7 +349,7 @@ async def test_get_streaming_no_fcc_in_response(
        model=openai_unit_test_env["OPENAI_CHAT_MODEL_ID"],
        stream=True,
        stream_options={"include_usage": True},
-        messages=openai_chat_completion._prepare_chat_history_for_request(orig_chat_history),  # type: ignore
+        messages=openai_chat_completion._prepare_messages_for_openai(orig_chat_history),  # type: ignore
    )


@@ -399,7 +399,7 @@ def test_chat_response_created_at_uses_utc(openai_unit_test_env: dict[str, str])
    )

    client = OpenAIChatClient()
-    response = client._create_chat_response(mock_response, ChatOptions())
+    response = client._parse_response_from_openai(mock_response, ChatOptions())

    # Verify that created_at is correctly formatted as UTC
    assert response.created_at is not None
@@ -431,7 +431,7 @@ def test_chat_response_update_created_at_uses_utc(openai_unit_test_env: dict[str
    )

    client = OpenAIChatClient()
-    response_update = client._create_chat_response_update(mock_chunk)
+    response_update = client._parse_response_update_from_openai(mock_chunk)

    # Verify that created_at is correctly formatted as UTC
    assert response_update.created_at is not None
@@ -368,16 +368,43 @@ async def test_response_format_parse_path() -> None:
    mock_parsed_response.output_parsed = None
    mock_parsed_response.usage = None
    mock_parsed_response.finish_reason = None
+    mock_parsed_response.conversation = None  # No conversation object

    with patch.object(client.client.responses, "parse", return_value=mock_parsed_response):
        response = await client.get_response(
            messages=[ChatMessage(role="user", text="Test message")], response_format=OutputStruct, store=True
        )
-
+        assert response.response_id == "parsed_response_123"
        assert response.conversation_id == "parsed_response_123"
        assert response.model_id == "test-model"


+async def test_response_format_parse_path_with_conversation_id() -> None:
+    """Test get_response response_format parsing path with set conversation ID."""
+    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")
+
+    # Mock successful parse response
+    mock_parsed_response = MagicMock()
+    mock_parsed_response.id = "parsed_response_123"
+    mock_parsed_response.text = "Parsed response"
+    mock_parsed_response.model = "test-model"
+    mock_parsed_response.created_at = 1000000000
+    mock_parsed_response.metadata = {}
+    mock_parsed_response.output_parsed = None
+    mock_parsed_response.usage = None
+    mock_parsed_response.finish_reason = None
+    mock_parsed_response.conversation = MagicMock()
+    mock_parsed_response.conversation.id = "conversation_456"
+
+    with patch.object(client.client.responses, "parse", return_value=mock_parsed_response):
+        response = await client.get_response(
+            messages=[ChatMessage(role="user", text="Test message")], response_format=OutputStruct, store=True
+        )
+        assert response.response_id == "parsed_response_123"
+        assert response.conversation_id == "conversation_456"
+        assert response.model_id == "test-model"
+
+
 async def test_bad_request_error_non_content_filter() -> None:
    """Test get_response BadRequestError without content_filter."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")
@@ -454,7 +481,7 @@ async def test_get_streaming_response_with_all_parameters() -> None:


 def test_response_content_creation_with_annotations() -> None:
-    """Test _create_response_content with different annotation types."""
+    """Test _parse_response_from_openai with different annotation types."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

    # Create a mock response with annotated text content
@@ -485,7 +512,7 @@ def test_response_content_creation_with_annotations() -> None:
    mock_response.output = [mock_message_item]

    with patch.object(client, "_get_metadata_from_response", return_value={}):
-        response = client._create_response_content(mock_response, chat_options=ChatOptions())  # type: ignore
+        response = client._parse_response_from_openai(mock_response, chat_options=ChatOptions())  # type: ignore

        assert len(response.messages[0].contents) >= 1
        assert isinstance(response.messages[0].contents[0], TextContent)
@@ -494,7 +521,7 @@ def test_response_content_creation_with_annotations() -> None:


 def test_response_content_creation_with_refusal() -> None:
-    """Test _create_response_content with refusal content."""
+    """Test _parse_response_from_openai with refusal content."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

    # Create a mock response with refusal content
@@ -516,7 +543,7 @@ def test_response_content_creation_with_refusal() -> None:

    mock_response.output = [mock_message_item]

-    response = client._create_response_content(mock_response, chat_options=ChatOptions())  # type: ignore
+    response = client._parse_response_from_openai(mock_response, chat_options=ChatOptions())  # type: ignore

    assert len(response.messages[0].contents) == 1
    assert isinstance(response.messages[0].contents[0], TextContent)
@@ -524,7 +551,7 @@ def test_response_content_creation_with_refusal() -> None:


 def test_response_content_creation_with_reasoning() -> None:
-    """Test _create_response_content with reasoning content."""
+    """Test _parse_response_from_openai with reasoning content."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

    # Create a mock response with reasoning content
@@ -546,7 +573,7 @@ def test_response_content_creation_with_reasoning() -> None:

    mock_response.output = [mock_reasoning_item]

-    response = client._create_response_content(mock_response, chat_options=ChatOptions())  # type: ignore
+    response = client._parse_response_from_openai(mock_response, chat_options=ChatOptions())  # type: ignore

    assert len(response.messages[0].contents) == 2
    assert isinstance(response.messages[0].contents[0], TextReasoningContent)
@@ -554,7 +581,7 @@ def test_response_content_creation_with_reasoning() -> None:


 def test_response_content_creation_with_code_interpreter() -> None:
-    """Test _create_response_content with code interpreter outputs."""
+    """Test _parse_response_from_openai with code interpreter outputs."""

    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

@@ -582,7 +609,7 @@ def test_response_content_creation_with_code_interpreter() -> None:

    mock_response.output = [mock_code_interpreter_item]

-    response = client._create_response_content(mock_response, chat_options=ChatOptions())  # type: ignore
+    response = client._parse_response_from_openai(mock_response, chat_options=ChatOptions())  # type: ignore

    assert len(response.messages[0].contents) == 2
    assert isinstance(response.messages[0].contents[0], TextContent)
@@ -593,7 +620,7 @@ def test_response_content_creation_with_code_interpreter() -> None:


 def test_response_content_creation_with_function_call() -> None:
-    """Test _create_response_content with function call content."""
+    """Test _parse_response_from_openai with function call content."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

    # Create a mock response with function call
@@ -614,7 +641,7 @@ def test_response_content_creation_with_function_call() -> None:

    mock_response.output = [mock_function_call_item]

-    response = client._create_response_content(mock_response, chat_options=ChatOptions())  # type: ignore
+    response = client._parse_response_from_openai(mock_response, chat_options=ChatOptions())  # type: ignore

    assert len(response.messages[0].contents) == 1
    assert isinstance(response.messages[0].contents[0], FunctionCallContent)
@@ -624,7 +651,7 @@ def test_response_content_creation_with_function_call() -> None:
    assert function_call.arguments == '{"location": "Seattle"}'


-def test_tools_to_response_tools_with_hosted_mcp() -> None:
+def test_prepare_tools_for_openai_with_hosted_mcp() -> None:
    """Test that HostedMCPTool is converted to the correct response tool dict."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

@@ -638,7 +665,7 @@ def test_tools_to_response_tools_with_hosted_mcp() -> None:
        additional_properties={"custom": "value"},
    )

-    resp_tools = client._tools_to_response_tools([tool])
+    resp_tools = client._prepare_tools_for_openai([tool])
    assert isinstance(resp_tools, list)
    assert len(resp_tools) == 1
    mcp = resp_tools[0]
@@ -654,7 +681,7 @@ def test_tools_to_response_tools_with_hosted_mcp() -> None:
    assert "require_approval" in mcp


-def test_create_response_content_with_mcp_approval_request() -> None:
+def test_parse_response_from_openai_with_mcp_approval_request() -> None:
    """Test that a non-streaming mcp_approval_request is parsed into FunctionApprovalRequestContent."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

@@ -675,7 +702,7 @@ def test_create_response_content_with_mcp_approval_request() -> None:

    mock_response.output = [mock_item]

-    response = client._create_response_content(mock_response, chat_options=ChatOptions())  # type: ignore
+    response = client._parse_response_from_openai(mock_response, chat_options=ChatOptions())  # type: ignore

    assert isinstance(response.messages[0].contents[0], FunctionApprovalRequestContent)
    req = response.messages[0].contents[0]
@@ -716,7 +743,7 @@ def test_responses_client_created_at_uses_utc(openai_unit_test_env: dict[str, st
    mock_response.output = [mock_message_item]

    with patch.object(client, "_get_metadata_from_response", return_value={}):
-        response = client._create_response_content(mock_response, chat_options=ChatOptions())  # type: ignore
+        response = client._parse_response_from_openai(mock_response, chat_options=ChatOptions())  # type: ignore

    # Verify that created_at is correctly formatted as UTC
    assert response.created_at is not None
@@ -730,7 +757,7 @@ def test_responses_client_created_at_uses_utc(openai_unit_test_env: dict[str, st
    )


-def test_tools_to_response_tools_with_raw_image_generation() -> None:
+def test_prepare_tools_for_openai_with_raw_image_generation() -> None:
    """Test that raw image_generation tool dict is handled correctly with parameter mapping."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

@@ -744,7 +771,7 @@ def test_tools_to_response_tools_with_raw_image_generation() -> None:
        "background": "transparent",
    }

-    resp_tools = client._tools_to_response_tools([tool])
+    resp_tools = client._prepare_tools_for_openai([tool])
    assert isinstance(resp_tools, list)
    assert len(resp_tools) == 1

@@ -759,7 +786,7 @@ def test_tools_to_response_tools_with_raw_image_generation() -> None:
    assert image_tool["output_compression"] == 75


-def test_tools_to_response_tools_with_raw_image_generation_openai_responses_params() -> None:
+def test_prepare_tools_for_openai_with_raw_image_generation_openai_responses_params() -> None:
    """Test raw image_generation tool with OpenAI-specific parameters."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

@@ -773,7 +800,7 @@ def test_tools_to_response_tools_with_raw_image_generation_openai_responses_para
        "partial_images": 2,  # Should be integer 0-3
    }

-    resp_tools = client._tools_to_response_tools([tool])
+    resp_tools = client._prepare_tools_for_openai([tool])
    assert isinstance(resp_tools, list)
    assert len(resp_tools) == 1

@@ -791,14 +818,14 @@ def test_tools_to_response_tools_with_raw_image_generation_openai_responses_para
    assert tool_dict["partial_images"] == 2


-def test_tools_to_response_tools_with_raw_image_generation_minimal() -> None:
+def test_prepare_tools_for_openai_with_raw_image_generation_minimal() -> None:
    """Test raw image_generation tool with minimal configuration."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

    # Test with minimal parameters (just type)
    tool = {"type": "image_generation"}

-    resp_tools = client._tools_to_response_tools([tool])
+    resp_tools = client._prepare_tools_for_openai([tool])
    assert isinstance(resp_tools, list)
    assert len(resp_tools) == 1

@@ -809,7 +836,7 @@ def test_tools_to_response_tools_with_raw_image_generation_minimal() -> None:
    assert len(image_tool) == 1


-def test_create_streaming_response_content_with_mcp_approval_request() -> None:
+def test_parse_chunk_from_openai_with_mcp_approval_request() -> None:
    """Test that a streaming mcp_approval_request event is parsed into FunctionApprovalRequestContent."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")
    chat_options = ChatOptions()
@@ -825,7 +852,7 @@ def test_create_streaming_response_content_with_mcp_approval_request() -> None:
    mock_item.server_label = "My_MCP"
    mock_event.item = mock_item

-    update = client._create_streaming_response_content(mock_event, chat_options, function_call_ids)
+    update = client._parse_chunk_from_openai(mock_event, chat_options, function_call_ids)
    assert any(isinstance(c, FunctionApprovalRequestContent) for c in update.contents)
    fa = next(c for c in update.contents if isinstance(c, FunctionApprovalRequestContent))
    assert fa.id == "approval-stream-1"
@@ -901,7 +928,7 @@ async def test_end_to_end_mcp_approval_flow(span_exporter) -> None:


 def test_usage_details_basic() -> None:
-    """Test _usage_details_from_openai without cached or reasoning tokens."""
+    """Test _parse_usage_from_openai without cached or reasoning tokens."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

    mock_usage = MagicMock()
@@ -911,7 +938,7 @@ def test_usage_details_basic() -> None:
    mock_usage.input_tokens_details = None
    mock_usage.output_tokens_details = None

-    details = client._usage_details_from_openai(mock_usage)  # type: ignore
+    details = client._parse_usage_from_openai(mock_usage)  # type: ignore
    assert details is not None
    assert details.input_token_count == 100
    assert details.output_token_count == 50
@@ -919,7 +946,7 @@ def test_usage_details_basic() -> None:


 def test_usage_details_with_cached_tokens() -> None:
-    """Test _usage_details_from_openai with cached input tokens."""
+    """Test _parse_usage_from_openai with cached input tokens."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

    mock_usage = MagicMock()
@@ -930,14 +957,14 @@ def test_usage_details_with_cached_tokens() -> None:
    mock_usage.input_tokens_details.cached_tokens = 25
    mock_usage.output_tokens_details = None

-    details = client._usage_details_from_openai(mock_usage)  # type: ignore
+    details = client._parse_usage_from_openai(mock_usage)  # type: ignore
    assert details is not None
    assert details.input_token_count == 200
    assert details.additional_counts["openai.cached_input_tokens"] == 25


 def test_usage_details_with_reasoning_tokens() -> None:
-    """Test _usage_details_from_openai with reasoning tokens."""
+    """Test _parse_usage_from_openai with reasoning tokens."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

    mock_usage = MagicMock()
@@ -948,7 +975,7 @@ def test_usage_details_with_reasoning_tokens() -> None:
    mock_usage.output_tokens_details = MagicMock()
    mock_usage.output_tokens_details.reasoning_tokens = 30

-    details = client._usage_details_from_openai(mock_usage)  # type: ignore
+    details = client._parse_usage_from_openai(mock_usage)  # type: ignore
    assert details is not None
    assert details.output_token_count == 80
    assert details.additional_counts["openai.reasoning_tokens"] == 30
@@ -975,7 +1002,7 @@ def test_get_metadata_from_response() -> None:


 def test_streaming_response_basic_structure() -> None:
-    """Test that _create_streaming_response_content returns proper structure."""
+    """Test that _parse_chunk_from_openai returns proper structure."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")
    chat_options = ChatOptions(store=True)
    function_call_ids: dict[int, tuple[str, str]] = {}
@@ -983,7 +1010,7 @@ def test_streaming_response_basic_structure() -> None:
    # Test with a basic mock event to ensure the method returns proper structure
    mock_event = MagicMock()

-    response = client._create_streaming_response_content(mock_event, chat_options, function_call_ids)  # type: ignore
+    response = client._parse_chunk_from_openai(mock_event, chat_options, function_call_ids)  # type: ignore

    # Should get a valid ChatResponseUpdate structure
    assert isinstance(response, ChatResponseUpdate)
@@ -993,6 +1020,44 @@ def test_streaming_response_basic_structure() -> None:
    assert response.raw_representation is mock_event


+def test_streaming_response_created_type() -> None:
+    """Test streaming response with created type"""
+    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")
+    chat_options = ChatOptions()
+    function_call_ids: dict[int, tuple[str, str]] = {}
+
+    mock_event = MagicMock()
+    mock_event.type = "response.created"
+    mock_event.response = MagicMock()
+    mock_event.response.id = "resp_1234"
+    mock_event.response.conversation = MagicMock()
+    mock_event.response.conversation.id = "conv_5678"
+
+    response = client._parse_chunk_from_openai(mock_event, chat_options, function_call_ids)
+
+    assert response.response_id == "resp_1234"
+    assert response.conversation_id == "conv_5678"
+
+
+def test_streaming_response_in_progress_type() -> None:
+    """Test streaming response with in_progress type"""
+    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")
+    chat_options = ChatOptions()
+    function_call_ids: dict[int, tuple[str, str]] = {}
+
+    mock_event = MagicMock()
+    mock_event.type = "response.in_progress"
+    mock_event.response = MagicMock()
+    mock_event.response.id = "resp_1234"
+    mock_event.response.conversation = MagicMock()
+    mock_event.response.conversation.id = "conv_5678"
+
+    response = client._parse_chunk_from_openai(mock_event, chat_options, function_call_ids)
+
+    assert response.response_id == "resp_1234"
+    assert response.conversation_id == "conv_5678"
+
+
 def test_streaming_annotation_added_with_file_path() -> None:
    """Test streaming annotation added event with file_path type extracts HostedFileContent."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")
@@ -1008,7 +1073,7 @@ def test_streaming_annotation_added_with_file_path() -> None:
        "index": 42,
    }

-    response = client._create_streaming_response_content(mock_event, chat_options, function_call_ids)
+    response = client._parse_chunk_from_openai(mock_event, chat_options, function_call_ids)

    assert len(response.contents) == 1
    content = response.contents[0]
@@ -1035,7 +1100,7 @@ def test_streaming_annotation_added_with_file_citation() -> None:
        "index": 15,
    }

-    response = client._create_streaming_response_content(mock_event, chat_options, function_call_ids)
+    response = client._parse_chunk_from_openai(mock_event, chat_options, function_call_ids)

    assert len(response.contents) == 1
    content = response.contents[0]
@@ -1064,7 +1129,7 @@ def test_streaming_annotation_added_with_container_file_citation() -> None:
        "end_index": 50,
    }

-    response = client._create_streaming_response_content(mock_event, chat_options, function_call_ids)
+    response = client._parse_chunk_from_openai(mock_event, chat_options, function_call_ids)

    assert len(response.contents) == 1
    content = response.contents[0]
@@ -1091,7 +1156,7 @@ def test_streaming_annotation_added_with_unknown_type() -> None:
        "url": "https://example.com",
    }

-    response = client._create_streaming_response_content(mock_event, chat_options, function_call_ids)
+    response = client._parse_chunk_from_openai(mock_event, chat_options, function_call_ids)

    # url_citation should not produce HostedFileContent
    assert len(response.contents) == 0
@@ -1137,8 +1202,8 @@ def test_get_streaming_response_with_response_format() -> None:
        asyncio.run(run_streaming())


-def test_openai_content_parser_image_content() -> None:
-    """Test _openai_content_parser with image content variations."""
+def test_prepare_content_for_openai_image_content() -> None:
+    """Test _prepare_content_for_openai with image content variations."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

    # Test image content with detail parameter and file_id
@@ -1147,7 +1212,7 @@ def test_openai_content_parser_image_content() -> None:
        media_type="image/jpeg",
        additional_properties={"detail": "high", "file_id": "file_123"},
    )
-    result = client._openai_content_parser(Role.USER, image_content_with_detail, {})  # type: ignore
+    result = client._prepare_content_for_openai(Role.USER, image_content_with_detail, {})  # type: ignore
    assert result["type"] == "input_image"
    assert result["image_url"] == "https://example.com/image.jpg"
    assert result["detail"] == "high"
@@ -1155,47 +1220,47 @@ def test_openai_content_parser_image_content() -> None:

    # Test image content without additional properties (defaults)
    image_content_basic = UriContent(uri="https://example.com/basic.png", media_type="image/png")
-    result = client._openai_content_parser(Role.USER, image_content_basic, {})  # type: ignore
+    result = client._prepare_content_for_openai(Role.USER, image_content_basic, {})  # type: ignore
    assert result["type"] == "input_image"
    assert result["detail"] == "auto"
    assert result["file_id"] is None


-def test_openai_content_parser_audio_content() -> None:
-    """Test _openai_content_parser with audio content variations."""
+def test_prepare_content_for_openai_audio_content() -> None:
+    """Test _prepare_content_for_openai with audio content variations."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

    # Test WAV audio content
    wav_content = UriContent(uri="data:audio/wav;base64,abc123", media_type="audio/wav")
-    result = client._openai_content_parser(Role.USER, wav_content, {})  # type: ignore
+    result = client._prepare_content_for_openai(Role.USER, wav_content, {})  # type: ignore
    assert result["type"] == "input_audio"
    assert result["input_audio"]["data"] == "data:audio/wav;base64,abc123"
    assert result["input_audio"]["format"] == "wav"

    # Test MP3 audio content
    mp3_content = UriContent(uri="data:audio/mp3;base64,def456", media_type="audio/mp3")
-    result = client._openai_content_parser(Role.USER, mp3_content, {})  # type: ignore
+    result = client._prepare_content_for_openai(Role.USER, mp3_content, {})  # type: ignore
    assert result["type"] == "input_audio"
    assert result["input_audio"]["format"] == "mp3"


-def test_openai_content_parser_unsupported_content() -> None:
-    """Test _openai_content_parser with unsupported content types."""
+def test_prepare_content_for_openai_unsupported_content() -> None:
+    """Test _prepare_content_for_openai with unsupported content types."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

    # Test unsupported audio format
    unsupported_audio = UriContent(uri="data:audio/ogg;base64,ghi789", media_type="audio/ogg")
-    result = client._openai_content_parser(Role.USER, unsupported_audio, {})  # type: ignore
+    result = client._prepare_content_for_openai(Role.USER, unsupported_audio, {})  # type: ignore
    assert result == {}

    # Test non-media content
    text_uri_content = UriContent(uri="https://example.com/document.txt", media_type="text/plain")
-    result = client._openai_content_parser(Role.USER, text_uri_content, {})  # type: ignore
+    result = client._prepare_content_for_openai(Role.USER, text_uri_content, {})  # type: ignore
    assert result == {}


-def test_create_streaming_response_content_code_interpreter() -> None:
-    """Test _create_streaming_response_content with code_interpreter_call."""
+def test_parse_chunk_from_openai_code_interpreter() -> None:
+    """Test _parse_chunk_from_openai with code_interpreter_call."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")
    chat_options = ChatOptions()
    function_call_ids: dict[int, tuple[str, str]] = {}
@@ -1211,15 +1276,15 @@ def test_create_streaming_response_content_code_interpreter() -> None:
    mock_item_image.code = None
    mock_event_image.item = mock_item_image

-    result = client._create_streaming_response_content(mock_event_image, chat_options, function_call_ids)  # type: ignore
+    result = client._parse_chunk_from_openai(mock_event_image, chat_options, function_call_ids)  # type: ignore
    assert len(result.contents) == 1
    assert isinstance(result.contents[0], UriContent)
    assert result.contents[0].uri == "https://example.com/plot.png"
    assert result.contents[0].media_type == "image"


-def test_create_streaming_response_content_reasoning() -> None:
-    """Test _create_streaming_response_content with reasoning content."""
+def test_parse_chunk_from_openai_reasoning() -> None:
+    """Test _parse_chunk_from_openai with reasoning content."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")
    chat_options = ChatOptions()
    function_call_ids: dict[int, tuple[str, str]] = {}
@@ -1234,7 +1299,7 @@ def test_create_streaming_response_content_reasoning() -> None:
    mock_item_reasoning.summary = ["Problem analysis summary"]
    mock_event_reasoning.item = mock_item_reasoning

-    result = client._create_streaming_response_content(mock_event_reasoning, chat_options, function_call_ids)  # type: ignore
+    result = client._parse_chunk_from_openai(mock_event_reasoning, chat_options, function_call_ids)  # type: ignore
    assert len(result.contents) == 1
    assert isinstance(result.contents[0], TextReasoningContent)
    assert result.contents[0].text == "Analyzing the problem step by step..."
@@ -1242,8 +1307,8 @@ def test_create_streaming_response_content_reasoning() -> None:
        assert result.contents[0].additional_properties["summary"] == "Problem analysis summary"


-def test_openai_content_parser_text_reasoning_comprehensive() -> None:
-    """Test _openai_content_parser with TextReasoningContent all additional properties."""
+def test_prepare_content_for_openai_text_reasoning_comprehensive() -> None:
+    """Test _prepare_content_for_openai with TextReasoningContent all additional properties."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

    # Test TextReasoningContent with all additional properties
@@ -1255,7 +1320,7 @@ def test_openai_content_parser_text_reasoning_comprehensive() -> None:
            "encrypted_content": "secure_data_456",
        },
    )
-    result = client._openai_content_parser(Role.ASSISTANT, comprehensive_reasoning, {})  # type: ignore
+    result = client._prepare_content_for_openai(Role.ASSISTANT, comprehensive_reasoning, {})  # type: ignore
    assert result["type"] == "reasoning"
    assert result["summary"]["text"] == "Comprehensive reasoning summary"
    assert result["status"] == "in_progress"
@@ -1280,7 +1345,7 @@ def test_streaming_reasoning_text_delta_event() -> None:
    )

    with patch.object(client, "_get_metadata_from_response", return_value={}) as mock_metadata:
-        response = client._create_streaming_response_content(event, chat_options, function_call_ids)  # type: ignore
+        response = client._parse_chunk_from_openai(event, chat_options, function_call_ids)  # type: ignore

        assert len(response.contents) == 1
        assert isinstance(response.contents[0], TextReasoningContent)
@@ -1305,7 +1370,7 @@ def test_streaming_reasoning_text_done_event() -> None:
    )

    with patch.object(client, "_get_metadata_from_response", return_value={"test": "data"}) as mock_metadata:
-        response = client._create_streaming_response_content(event, chat_options, function_call_ids)  # type: ignore
+        response = client._parse_chunk_from_openai(event, chat_options, function_call_ids)  # type: ignore

        assert len(response.contents) == 1
        assert isinstance(response.contents[0], TextReasoningContent)
@@ -1331,7 +1396,7 @@ def test_streaming_reasoning_summary_text_delta_event() -> None:
    )

    with patch.object(client, "_get_metadata_from_response", return_value={}) as mock_metadata:
-        response = client._create_streaming_response_content(event, chat_options, function_call_ids)  # type: ignore
+        response = client._parse_chunk_from_openai(event, chat_options, function_call_ids)  # type: ignore

        assert len(response.contents) == 1
        assert isinstance(response.contents[0], TextReasoningContent)
@@ -1356,7 +1421,7 @@ def test_streaming_reasoning_summary_text_done_event() -> None:
    )

    with patch.object(client, "_get_metadata_from_response", return_value={"custom": "meta"}) as mock_metadata:
-        response = client._create_streaming_response_content(event, chat_options, function_call_ids)  # type: ignore
+        response = client._parse_chunk_from_openai(event, chat_options, function_call_ids)  # type: ignore

        assert len(response.contents) == 1
        assert isinstance(response.contents[0], TextReasoningContent)
@@ -1392,8 +1457,8 @@ def test_streaming_reasoning_events_preserve_metadata() -> None:
    )

    with patch.object(client, "_get_metadata_from_response", return_value={"test": "metadata"}):
-        text_response = client._create_streaming_response_content(text_event, chat_options, function_call_ids)  # type: ignore
-        reasoning_response = client._create_streaming_response_content(reasoning_event, chat_options, function_call_ids)  # type: ignore
+        text_response = client._parse_chunk_from_openai(text_event, chat_options, function_call_ids)  # type: ignore
+        reasoning_response = client._parse_chunk_from_openai(reasoning_event, chat_options, function_call_ids)  # type: ignore

        # Both should preserve metadata
        assert text_response.additional_properties == {"test": "metadata"}
@@ -1404,7 +1469,7 @@ def test_streaming_reasoning_events_preserve_metadata() -> None:
        assert isinstance(reasoning_response.contents[0], TextReasoningContent)


-def test_create_response_content_image_generation_raw_base64():
+def test_parse_response_from_openai_image_generation_raw_base64():
    """Test image generation response parsing with raw base64 string."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

@@ -1428,7 +1493,7 @@ def test_create_response_content_image_generation_raw_base64():
    mock_response.output = [mock_item]

    with patch.object(client, "_get_metadata_from_response", return_value={}):
-        response = client._create_response_content(mock_response, chat_options=ChatOptions())  # type: ignore
+        response = client._parse_response_from_openai(mock_response, chat_options=ChatOptions())  # type: ignore

    # Verify the response contains DataContent with proper URI and media_type
    assert len(response.messages[0].contents) == 1
@@ -1438,7 +1503,7 @@ def test_create_response_content_image_generation_raw_base64():
    assert content.media_type == "image/png"


-def test_create_response_content_image_generation_existing_data_uri():
+def test_parse_response_from_openai_image_generation_existing_data_uri():
    """Test image generation response parsing with existing data URI."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

@@ -1461,7 +1526,7 @@ def test_create_response_content_image_generation_existing_data_uri():
    mock_response.output = [mock_item]

    with patch.object(client, "_get_metadata_from_response", return_value={}):
-        response = client._create_response_content(mock_response, chat_options=ChatOptions())  # type: ignore
+        response = client._parse_response_from_openai(mock_response, chat_options=ChatOptions())  # type: ignore

    # Verify the response contains DataContent with proper media_type parsed from URI
    assert len(response.messages[0].contents) == 1
@@ -1471,7 +1536,7 @@ def test_create_response_content_image_generation_existing_data_uri():
    assert content.media_type == "image/webp"


-def test_create_response_content_image_generation_format_detection():
+def test_parse_response_from_openai_image_generation_format_detection():
    """Test different image format detection from base64 data."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

@@ -1493,7 +1558,7 @@ def test_create_response_content_image_generation_format_detection():
    mock_response_jpeg.output = [mock_item_jpeg]

    with patch.object(client, "_get_metadata_from_response", return_value={}):
-        response_jpeg = client._create_response_content(mock_response_jpeg, chat_options=ChatOptions())  # type: ignore
+        response_jpeg = client._parse_response_from_openai(mock_response_jpeg, chat_options=ChatOptions())  # type: ignore
    content_jpeg = response_jpeg.messages[0].contents[0]
    assert isinstance(content_jpeg, DataContent)
    assert content_jpeg.media_type == "image/jpeg"
@@ -1517,14 +1582,14 @@ def test_create_response_content_image_generation_format_detection():
    mock_response_webp.output = [mock_item_webp]

    with patch.object(client, "_get_metadata_from_response", return_value={}):
-        response_webp = client._create_response_content(mock_response_webp, chat_options=ChatOptions())  # type: ignore
+        response_webp = client._parse_response_from_openai(mock_response_webp, chat_options=ChatOptions())  # type: ignore
    content_webp = response_webp.messages[0].contents[0]
    assert isinstance(content_webp, DataContent)
    assert content_webp.media_type == "image/webp"
    assert "data:image/webp;base64," in content_webp.uri


-def test_create_response_content_image_generation_fallback():
+def test_parse_response_from_openai_image_generation_fallback():
    """Test image generation with invalid base64 falls back to PNG."""
    client = OpenAIResponsesClient(model_id="test-model", api_key="test-key")

@@ -1547,7 +1612,7 @@ def test_create_response_content_image_generation_fallback():
    mock_response.output = [mock_item]

    with patch.object(client, "_get_metadata_from_response", return_value={}):
-        response = client._create_response_content(mock_response, chat_options=ChatOptions())  # type: ignore
+        response = client._parse_response_from_openai(mock_response, chat_options=ChatOptions())  # type: ignore

    # Verify it falls back to PNG format for unrecognized binary data
    assert len(response.messages[0].contents) == 1
@@ -1563,21 +1628,21 @@ async def test_prepare_options_store_parameter_handling() -> None:

    test_conversation_id = "test-conversation-123"
    chat_options = ChatOptions(store=True, conversation_id=test_conversation_id)
-    options = await client.prepare_options(messages, chat_options)
+    options = await client._prepare_options(messages, chat_options)  # type: ignore
    assert options["store"] is True
    assert options["previous_response_id"] == test_conversation_id

    chat_options = ChatOptions(store=False, conversation_id="")
-    options = await client.prepare_options(messages, chat_options)
+    options = await client._prepare_options(messages, chat_options)  # type: ignore
    assert options["store"] is False

    chat_options = ChatOptions(store=None, conversation_id=None)
-    options = await client.prepare_options(messages, chat_options)
+    options = await client._prepare_options(messages, chat_options)  # type: ignore
    assert "store" not in options
    assert "previous_response_id" not in options

    chat_options = ChatOptions()
-    options = await client.prepare_options(messages, chat_options)
+    options = await client._prepare_options(messages, chat_options)  # type: ignore
    assert "store" not in options
    assert "previous_response_id" not in options

@@ -1,11 +1,13 @@
 # Copyright (c) Microsoft. All rights reserved.

 import uuid
+from collections.abc import AsyncIterable
 from typing import Any

 import pytest

 from agent_framework import (
+    AgentProtocol,
    AgentRunResponse,
    AgentRunResponseUpdate,
    AgentRunUpdateEvent,
@@ -422,6 +424,48 @@ class TestWorkflowAgent:
        assert isinstance(updates[2].raw_representation, CustomData)
        assert updates[2].raw_representation.value == 42

+    async def test_workflow_as_agent_yield_output_with_list_of_chat_messages(self) -> None:
+        """Test that yield_output with list[ChatMessage] extracts contents from all messages.
+
+        Note: TextContent items are coalesced by _finalize_response, so multiple text contents
+        become a single merged TextContent in the final response.
+        """
+
+        @executor
+        async def list_yielding_executor(messages: list[ChatMessage], ctx: WorkflowContext) -> None:
+            # Yield a list of ChatMessages (as SequentialBuilder does)
+            msg_list = [
+                ChatMessage(role=Role.USER, contents=[TextContent(text="first message")]),
+                ChatMessage(role=Role.ASSISTANT, contents=[TextContent(text="second message")]),
+                ChatMessage(
+                    role=Role.ASSISTANT,
+                    contents=[TextContent(text="third"), TextContent(text="fourth")],
+                ),
+            ]
+            await ctx.yield_output(msg_list)
+
+        workflow = WorkflowBuilder().set_start_executor(list_yielding_executor).build()
+        agent = workflow.as_agent("list-msg-agent")
+
+        # Verify streaming returns the update with all 4 contents before coalescing
+        updates: list[AgentRunResponseUpdate] = []
+        async for update in agent.run_stream("test"):
+            updates.append(update)
+
+        assert len(updates) == 1
+        assert len(updates[0].contents) == 4
+        texts = [c.text for c in updates[0].contents if isinstance(c, TextContent)]
+        assert texts == ["first message", "second message", "third", "fourth"]
+
+        # Verify run() coalesces text contents (expected behavior)
+        result = await agent.run("test")
+
+        assert isinstance(result, AgentRunResponse)
+        assert len(result.messages) == 1
+        # TextContent items are coalesced into one
+        assert len(result.messages[0].contents) == 1
+        assert result.messages[0].text == "first messagesecond messagethirdfourth"
+
    async def test_thread_conversation_history_included_in_workflow_run(self) -> None:
        """Test that conversation history from thread is included when running WorkflowAgent.

@@ -521,6 +565,142 @@ class TestWorkflowAgent:
        checkpoints = await checkpoint_storage.list_checkpoints(workflow.id)
        assert len(checkpoints) > 0, "Checkpoints should have been created when checkpoint_storage is provided"

+    async def test_agent_executor_output_response_false_filters_streaming_events(self):
+        """Test that AgentExecutor with output_response=False does not surface streaming events."""
+
+        class MockAgent(AgentProtocol):
+            """Mock agent for testing."""
+
+            def __init__(self, name: str, response_text: str) -> None:
+                self._name = name
+                self._response_text = response_text
+                self._description: str | None = None
+
+            @property
+            def name(self) -> str | None:
+                return self._name
+
+            @property
+            def description(self) -> str | None:
+                return self._description
+
+            def get_new_thread(self) -> AgentThread:
+                return AgentThread()
+
+            async def run(self, messages: Any, *, thread: AgentThread | None = None, **kwargs: Any) -> AgentRunResponse:
+                return AgentRunResponse(
+                    messages=[ChatMessage(role=Role.ASSISTANT, text=self._response_text)],
+                    text=self._response_text,
+                )
+
+            async def run_stream(
+                self, messages: Any, *, thread: AgentThread | None = None, **kwargs: Any
+            ) -> AsyncIterable[AgentRunResponseUpdate]:
+                for word in self._response_text.split():
+                    yield AgentRunResponseUpdate(
+                        contents=[TextContent(text=word + " ")],
+                        role=Role.ASSISTANT,
+                        author_name=self._name,
+                    )
+
+        @executor
+        async def start_executor(messages: list[ChatMessage], ctx: WorkflowContext) -> None:
+            from agent_framework import AgentExecutorRequest
+
+            await ctx.yield_output("Start output")
+            await ctx.send_message(AgentExecutorRequest(messages=messages, should_respond=True))
+
+        # Build workflow: start -> agent1 (no output) -> agent2 (output_response=True)
+        workflow = (
+            WorkflowBuilder()
+            .register_executor(lambda: start_executor, "start")
+            .register_agent(lambda: MockAgent("agent1", "Agent1 output - should NOT appear"), "agent1")
+            .register_agent(
+                lambda: MockAgent("agent2", "Agent2 output - SHOULD appear"), "agent2", output_response=True
+            )
+            .set_start_executor("start")
+            .add_edge("start", "agent1")
+            .add_edge("agent1", "agent2")
+            .build()
+        )
+
+        agent = WorkflowAgent(workflow=workflow, name="Test Agent")
+        result = await agent.run("Test input")
+
+        # Collect all message texts
+        texts = [msg.text for msg in result.messages if msg.text]
+
+        # Start output should appear (from yield_output)
+        assert any("Start output" in t for t in texts), "Start output should appear"
+
+        # Agent1 output should NOT appear (output_response=False)
+        assert not any("Agent1" in t for t in texts), "Agent1 output should NOT appear"
+
+        # Agent2 output should appear (output_response=True)
+        assert any("Agent2" in t for t in texts), "Agent2 output should appear"
+
+    async def test_agent_executor_output_response_no_duplicate_from_workflow_output_event(self):
+        """Test that AgentExecutor with output_response=True does not duplicate content."""
+
+        class MockAgent(AgentProtocol):
+            """Mock agent for testing."""
+
+            def __init__(self, name: str, response_text: str) -> None:
+                self._name = name
+                self._response_text = response_text
+                self._description: str | None = None
+
+            @property
+            def name(self) -> str | None:
+                return self._name
+
+            @property
+            def description(self) -> str | None:
+                return self._description
+
+            def get_new_thread(self) -> AgentThread:
+                return AgentThread()
+
+            async def run(self, messages: Any, *, thread: AgentThread | None = None, **kwargs: Any) -> AgentRunResponse:
+                return AgentRunResponse(
+                    messages=[ChatMessage(role=Role.ASSISTANT, text=self._response_text)],
+                    text=self._response_text,
+                )
+
+            async def run_stream(
+                self, messages: Any, *, thread: AgentThread | None = None, **kwargs: Any
+            ) -> AsyncIterable[AgentRunResponseUpdate]:
+                yield AgentRunResponseUpdate(
+                    contents=[TextContent(text=self._response_text)],
+                    role=Role.ASSISTANT,
+                    author_name=self._name,
+                )
+
+        @executor
+        async def start_executor(messages: list[ChatMessage], ctx: WorkflowContext) -> None:
+            from agent_framework import AgentExecutorRequest
+
+            await ctx.send_message(AgentExecutorRequest(messages=messages, should_respond=True))
+
+        # Build workflow with single agent that has output_response=True
+        workflow = (
+            WorkflowBuilder()
+            .register_executor(lambda: start_executor, "start")
+            .register_agent(lambda: MockAgent("agent", "Unique response text"), "agent", output_response=True)
+            .set_start_executor("start")
+            .add_edge("start", "agent")
+            .build()
+        )
+
+        agent = WorkflowAgent(workflow=workflow, name="Test Agent")
+        result = await agent.run("Test input")
+
+        # Count occurrences of the unique response text
+        unique_text_count = sum(1 for msg in result.messages if msg.text and "Unique response text" in msg.text)
+
+        # Should appear exactly once (not duplicated from both streaming and WorkflowOutputEvent)
+        assert unique_text_count == 1, f"Response should appear exactly once, but appeared {unique_text_count} times"
+

 class TestWorkflowAgentMergeUpdates:
    """Test cases specifically for the WorkflowAgent.merge_updates static method."""
@@ -245,7 +245,8 @@ def test_register_multiple_executors():

    # Build workflow with edges using registered names
    workflow = (
-        builder.set_start_executor("ExecutorA")
+        builder
+        .set_start_executor("ExecutorA")
        .add_edge("ExecutorA", "ExecutorB")
        .add_edge("ExecutorB", "ExecutorC")
        .build()
@@ -426,7 +427,8 @@ def test_register_with_fan_in_edges():
    # Add fan-in edges using registered names
    # Both Source1 and Source2 need to be reachable, so connect Source1 to Source2
    workflow = (
-        builder.set_start_executor("Source1")
+        builder
+        .set_start_executor("Source1")
        .add_edge("Source1", "Source2")
        .add_fan_in_edges(["Source1", "Source2"], "Aggregator")
        .build()
@@ -490,3 +490,266 @@ async def test_magentic_kwargs_stored_in_shared_state() -> None:


 # endregion
+
+
+# region WorkflowAgent (as_agent) kwargs Tests
+
+
+async def test_workflow_as_agent_run_propagates_kwargs_to_underlying_agent() -> None:
+    """Test that kwargs passed to workflow_agent.run() flow through to the underlying agents."""
+    agent = _KwargsCapturingAgent(name="inner_agent")
+    workflow = SequentialBuilder().participants([agent]).build()
+    workflow_agent = workflow.as_agent(name="TestWorkflowAgent")
+
+    custom_data = {"endpoint": "https://api.example.com", "version": "v1"}
+    user_token = {"user_name": "alice", "access_level": "admin"}
+
+    _ = await workflow_agent.run(
+        "test message",
+        custom_data=custom_data,
+        user_token=user_token,
+    )
+
+    # Verify inner agent received kwargs
+    assert len(agent.captured_kwargs) >= 1, "Inner agent should have been invoked at least once"
+    received = agent.captured_kwargs[0]
+    assert "custom_data" in received, "Inner agent should receive custom_data kwarg"
+    assert "user_token" in received, "Inner agent should receive user_token kwarg"
+    assert received["custom_data"] == custom_data
+    assert received["user_token"] == user_token
+
+
+async def test_workflow_as_agent_run_stream_propagates_kwargs_to_underlying_agent() -> None:
+    """Test that kwargs passed to workflow_agent.run_stream() flow through to the underlying agents."""
+    agent = _KwargsCapturingAgent(name="inner_agent")
+    workflow = SequentialBuilder().participants([agent]).build()
+    workflow_agent = workflow.as_agent(name="TestWorkflowAgent")
+
+    custom_data = {"session_id": "xyz123"}
+    api_token = "secret-token"
+
+    async for _ in workflow_agent.run_stream(
+        "test message",
+        custom_data=custom_data,
+        api_token=api_token,
+    ):
+        pass
+
+    # Verify inner agent received kwargs
+    assert len(agent.captured_kwargs) >= 1, "Inner agent should have been invoked at least once"
+    received = agent.captured_kwargs[0]
+    assert "custom_data" in received, "Inner agent should receive custom_data kwarg"
+    assert "api_token" in received, "Inner agent should receive api_token kwarg"
+    assert received["custom_data"] == custom_data
+    assert received["api_token"] == api_token
+
+
+async def test_workflow_as_agent_propagates_kwargs_to_multiple_agents() -> None:
+    """Test that kwargs flow to all agents when using workflow.as_agent()."""
+    agent1 = _KwargsCapturingAgent(name="agent1")
+    agent2 = _KwargsCapturingAgent(name="agent2")
+    workflow = SequentialBuilder().participants([agent1, agent2]).build()
+    workflow_agent = workflow.as_agent(name="MultiAgentWorkflow")
+
+    custom_data = {"batch_id": "batch-001"}
+
+    _ = await workflow_agent.run("test message", custom_data=custom_data)
+
+    # Both agents should have received kwargs
+    assert len(agent1.captured_kwargs) >= 1, "First agent should be invoked"
+    assert len(agent2.captured_kwargs) >= 1, "Second agent should be invoked"
+    assert agent1.captured_kwargs[0].get("custom_data") == custom_data
+    assert agent2.captured_kwargs[0].get("custom_data") == custom_data
+
+
+async def test_workflow_as_agent_kwargs_with_none_values() -> None:
+    """Test that kwargs with None values are passed through correctly via as_agent()."""
+    agent = _KwargsCapturingAgent(name="none_test_agent")
+    workflow = SequentialBuilder().participants([agent]).build()
+    workflow_agent = workflow.as_agent(name="NoneTestWorkflow")
+
+    _ = await workflow_agent.run("test", optional_param=None, other_param="value")
+
+    assert len(agent.captured_kwargs) >= 1
+    received = agent.captured_kwargs[0]
+    assert "optional_param" in received
+    assert received["optional_param"] is None
+    assert received["other_param"] == "value"
+
+
+async def test_workflow_as_agent_kwargs_with_complex_nested_data() -> None:
+    """Test that complex nested data structures flow through correctly via as_agent()."""
+    agent = _KwargsCapturingAgent(name="nested_agent")
+    workflow = SequentialBuilder().participants([agent]).build()
+    workflow_agent = workflow.as_agent(name="NestedDataWorkflow")
+
+    complex_data = {
+        "level1": {
+            "level2": {
+                "level3": ["a", "b", "c"],
+                "number": 42,
+            },
+            "list": [1, 2, {"nested": True}],
+        },
+    }
+
+    _ = await workflow_agent.run("test", complex_data=complex_data)
+
+    assert len(agent.captured_kwargs) >= 1
+    received = agent.captured_kwargs[0]
+    assert received.get("complex_data") == complex_data
+
+
+# endregion
+
+
+# region SubWorkflow (WorkflowExecutor) Tests
+
+
+async def test_subworkflow_kwargs_propagation() -> None:
+    """Test that kwargs are propagated to subworkflows.
+
+    Verifies kwargs passed to parent workflow.run_stream() flow through to agents
+    in subworkflows wrapped by WorkflowExecutor.
+    """
+    from agent_framework._workflows._workflow_executor import WorkflowExecutor
+
+    # Create an agent inside the subworkflow that captures kwargs
+    inner_agent = _KwargsCapturingAgent(name="inner_agent")
+
+    # Build the inner (sub) workflow with the agent
+    inner_workflow = SequentialBuilder().participants([inner_agent]).build()
+
+    # Wrap the inner workflow in a WorkflowExecutor so it can be used as a subworkflow
+    subworkflow_executor = WorkflowExecutor(workflow=inner_workflow, id="subworkflow_executor")
+
+    # Build the outer (parent) workflow containing the subworkflow
+    outer_workflow = SequentialBuilder().participants([subworkflow_executor]).build()
+
+    # Define kwargs that should propagate to subworkflow
+    custom_data = {"api_key": "secret123", "endpoint": "https://api.example.com"}
+    user_token = {"user_name": "alice", "access_level": "admin"}
+
+    # Run the outer workflow with kwargs
+    async for event in outer_workflow.run_stream(
+        "test message for subworkflow",
+        custom_data=custom_data,
+        user_token=user_token,
+    ):
+        if isinstance(event, WorkflowStatusEvent) and event.state == WorkflowRunState.IDLE:
+            break
+
+    # Verify that the inner agent was called
+    assert len(inner_agent.captured_kwargs) >= 1, "Inner agent in subworkflow should have been invoked"
+
+    received_kwargs = inner_agent.captured_kwargs[0]
+
+    # Verify kwargs were propagated from parent workflow to subworkflow agent
+    assert "custom_data" in received_kwargs, (
+        f"Subworkflow agent should receive 'custom_data' kwarg. Received keys: {list(received_kwargs.keys())}"
+    )
+    assert "user_token" in received_kwargs, (
+        f"Subworkflow agent should receive 'user_token' kwarg. Received keys: {list(received_kwargs.keys())}"
+    )
+    assert received_kwargs.get("custom_data") == custom_data, (
+        f"Expected custom_data={custom_data}, got {received_kwargs.get('custom_data')}"
+    )
+    assert received_kwargs.get("user_token") == user_token, (
+        f"Expected user_token={user_token}, got {received_kwargs.get('user_token')}"
+    )
+
+
+async def test_subworkflow_kwargs_accessible_via_shared_state() -> None:
+    """Test that kwargs are accessible via SharedState within subworkflow.
+
+    Verifies that WORKFLOW_RUN_KWARGS_KEY is populated in the subworkflow's SharedState
+    with kwargs from the parent workflow.
+    """
+    from agent_framework import Executor, WorkflowContext, handler
+    from agent_framework._workflows._workflow_executor import WorkflowExecutor
+
+    captured_kwargs_from_state: list[dict[str, Any]] = []
+
+    class _SharedStateReader(Executor):
+        """Executor that reads kwargs from SharedState for verification."""
+
+        @handler
+        async def read_kwargs(self, msgs: list[ChatMessage], ctx: WorkflowContext[list[ChatMessage]]) -> None:
+            kwargs_from_state = await ctx.get_shared_state(WORKFLOW_RUN_KWARGS_KEY)
+            captured_kwargs_from_state.append(kwargs_from_state or {})
+            await ctx.send_message(msgs)
+
+    # Build inner workflow with SharedState reader
+    state_reader = _SharedStateReader(id="state_reader")
+    inner_workflow = SequentialBuilder().participants([state_reader]).build()
+
+    # Wrap as subworkflow
+    subworkflow_executor = WorkflowExecutor(workflow=inner_workflow, id="subworkflow")
+
+    # Build outer workflow
+    outer_workflow = SequentialBuilder().participants([subworkflow_executor]).build()
+
+    # Run with kwargs
+    async for event in outer_workflow.run_stream(
+        "test",
+        my_custom_kwarg="should_be_propagated",
+        another_kwarg=42,
+    ):
+        if isinstance(event, WorkflowStatusEvent) and event.state == WorkflowRunState.IDLE:
+            break
+
+    # Verify the state reader was invoked
+    assert len(captured_kwargs_from_state) >= 1, "SharedState reader should have been invoked"
+
+    kwargs_in_subworkflow = captured_kwargs_from_state[0]
+
+    assert kwargs_in_subworkflow.get("my_custom_kwarg") == "should_be_propagated", (
+        f"Expected 'my_custom_kwarg' in subworkflow SharedState, got: {kwargs_in_subworkflow}"
+    )
+    assert kwargs_in_subworkflow.get("another_kwarg") == 42, (
+        f"Expected 'another_kwarg'=42 in subworkflow SharedState, got: {kwargs_in_subworkflow}"
+    )
+
+
+async def test_nested_subworkflow_kwargs_propagation() -> None:
+    """Test kwargs propagation through multiple levels of nested subworkflows.
+
+    Verifies kwargs flow through 3 levels:
+    - Outer workflow
+      - Middle subworkflow (WorkflowExecutor)
+        - Inner subworkflow (WorkflowExecutor) with agent
+    """
+    from agent_framework._workflows._workflow_executor import WorkflowExecutor
+
+    # Innermost agent
+    inner_agent = _KwargsCapturingAgent(name="deeply_nested_agent")
+
+    # Build inner workflow
+    inner_workflow = SequentialBuilder().participants([inner_agent]).build()
+    inner_executor = WorkflowExecutor(workflow=inner_workflow, id="inner_executor")
+
+    # Build middle workflow containing inner
+    middle_workflow = SequentialBuilder().participants([inner_executor]).build()
+    middle_executor = WorkflowExecutor(workflow=middle_workflow, id="middle_executor")
+
+    # Build outer workflow containing middle
+    outer_workflow = SequentialBuilder().participants([middle_executor]).build()
+
+    # Run with kwargs
+    async for event in outer_workflow.run_stream(
+        "deeply nested test",
+        deep_kwarg="should_reach_inner",
+    ):
+        if isinstance(event, WorkflowStatusEvent) and event.state == WorkflowRunState.IDLE:
+            break
+
+    # Verify inner agent was called
+    assert len(inner_agent.captured_kwargs) >= 1, "Deeply nested agent should be invoked"
+
+    received = inner_agent.captured_kwargs[0]
+    assert received.get("deep_kwarg") == "should_reach_inner", (
+        f"Deeply nested agent should receive 'deep_kwarg'. Got: {received}"
+    )
+
+
+# endregion