Compare commits

...
Author SHA1 Message Date
Shyju Krishnankutty ad4b732741 Adding ReflectExecutors method to Workflow. 2026-01-22 12:07:03 -08:00
Dmytro StrukandGitHub b4a71f00a3 Updated package versions (#3335) 2026-01-21 18:40:36 +00:00
SukeeshandGitHub 082f39e77e Python: feat(anthropic): Add response_format support for structured outputs (#3301)
* fix(anthropic): Add response_format support for structured outputs

* only use from options

* use native way of response format

* ruff lint fix

* address comment; handle dict
2026-01-21 15:08:13 +00:00
CopilotGitHubstephentoubcopilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>Roger Barreto
6f1ab66795 .NET: Fix DebuggerDisplay attribute in AIAgent.cs to reference existing properties (#2985)
* Initial plan

* Fix DebuggerDisplay attribute in AIAgent.cs to reference existing properties

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>
Co-authored-by: Roger Barreto <19890735+rogerbarreto@users.noreply.github.com>
2026-01-21 13:50:59 +00:00
dependabot[bot]GitHubdependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
d402d92a47 Bump pyasn1 from 0.6.1 to 0.6.2 in /python (#3257)
Bumps [pyasn1](https://github.com/pyasn1/pyasn1) from 0.6.1 to 0.6.2.
- [Release notes](https://github.com/pyasn1/pyasn1/releases)
- [Changelog](https://github.com/pyasn1/pyasn1/blob/main/CHANGES.rst)
- [Commits](https://github.com/pyasn1/pyasn1/compare/v0.6.1...v0.6.2)

---
updated-dependencies:
- dependency-name: pyasn1
  dependency-version: 0.6.2
  dependency-type: indirect
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2026-01-21 12:56:40 +00:00
d55dd5f253 .NET: Improve readme for agents V2 (#3285)
* Improve readme for agents V2

* Architectural justification

* Update dotnet/samples/GettingStarted/FoundryAgents/README.md

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
2026-01-21 12:55:11 +00:00
Giles OdigweandGitHub 88e0ee1a2c Python: Fix local MCP tools with AzureAIProjectAgentProvider (#3315)
* azureai v2 local mcp fix

* addressed copilot comments
2026-01-21 12:43:05 +00:00
Roger BarretoandGitHub aa6579f38c .NET: Update Conversation Sample to use Conversation Id instead (#3180)
* Update Conversation Sample to use conversation Id instead

* Remove Run infix

* Remove the sync GetAIAgent from sample
2026-01-21 12:15:13 +00:00
41cc34421f .NET: Add sample to show multiple AIContextProvider usage (#3284)
* Add sample to show multiple AIContextProvider usage

* Update comment.

* Update messaging in README.

* Address PR comments.

---------

Co-authored-by: Mark Wallace <127216156+markwallace-microsoft@users.noreply.github.com>
2026-01-21 11:42:16 +00:00
CopilotGitHubrogerbarretocopilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
eac8baac09 .Net: Fix DebuggerDisplay attribute to reference existing property (#3326)
* Initial plan

* Fix DebuggerDisplay attribute to use Name instead of DisplayName

Co-authored-by: rogerbarreto <19890735+rogerbarreto@users.noreply.github.com>

---------

Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: rogerbarreto <19890735+rogerbarreto@users.noreply.github.com>
2026-01-21 11:42:01 +00:00
dependabot[bot]GitHubdependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
77236bf0ec Bump tomli from 2.3.0 to 2.4.0 in /python (#3182)
Bumps [tomli](https://github.com/hukkin/tomli) from 2.3.0 to 2.4.0.
- [Changelog](https://github.com/hukkin/tomli/blob/master/CHANGELOG.md)
- [Commits](https://github.com/hukkin/tomli/compare/2.3.0...2.4.0)

---
updated-dependencies:
- dependency-name: tomli
  dependency-version: 2.4.0
  dependency-type: direct:development
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2026-01-21 05:38:12 +00:00
Evan MattsonandGitHub 6d7690e485 Python: fix(ag-ui): properly handle json serialize with handoff workflows as agent (#3275)
* fix(ag-ui): properly handle json serialize with handoff workflows as agent

* Other improvements around handling non-serializable objects
2026-01-21 02:43:14 +00:00
Evan MattsonandGitHub 6b5437e4ec Python: fix(core): handle anyio cancel scope errors during MCP connection cleanup (#3277)
* fix(core): handle anyio cancel scope errors during MCP connection cleanup

* Address Copilot feedback
2026-01-21 01:49:07 +00:00
db8a59bd3d .NET: Durable Agent samples and automated validation for non-Azure Functions (#3042)
* Durable Agent samples and automated validation for non-Azure Functions

* Update test projects

* fix file encoding

* Remove AgentThreadMetadata usage

* Absorb breaking change from #3152

* Absorb newer breaking changes (AgentRunResponse --> AgentResponse)

* Absorb more breaking changes (see #3222)

* Improve integration test reliability (isolated task hubs, etc.)

* Fix flakey streaming test

---------

Co-authored-by: Mark Wallace <127216156+markwallace-microsoft@users.noreply.github.com>
2026-01-20 22:45:10 +00:00
Eduard van ValkenburgandGitHub 83e6229c11 Python: [Breaking] Simplified Content types to a single class with classmethod constructors. (#3252)
* ported Content to a new model

* fixed linting

* fixes

* fixed data format handling

* fix for 3.10 mypy

* fix

* fix int test
2026-01-20 22:09:39 +00:00
73761aa4a3 .NET: Pass AdditionalProperties from parent to child when exposing an agent as a FunctionTool (#3219)
* Pass AdditionalProperties from parent to child when exposing an agent as a FunctionTool

* Rename variable to improve readability.

* Apply suggestions from code review

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
2026-01-20 18:20:08 +00:00
CopilotGitHubstephentoubcopilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
742937194a .NET: Update Microsoft.Extensions.AI.* packages to 10.2.0 (#3211)
* Initial plan

* Update Microsoft.Extensions.AI.* to 10.2.0 and fix timestamp behavior tests

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>
2026-01-20 17:52:58 +00:00
74401266e6 Improve PR number handling in workflow (#3302)
* Improve PR number handling in workflow

Refine PR number extraction and validation method.

* Update .github/workflows/python-test-coverage-report.yml

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Fix error message for invalid PR number

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
2026-01-20 14:05:15 +00:00
f8c84d4ee6 Python: Fix: Add system_instructions to ChatClient LLM span tracing (#3164)
* Fix: Add system_instructions to ChatClient LLM span tracing

- Add system_instructions parameter to _capture_messages() calls in
  _trace_get_response() and _trace_get_streaming_response()
- Extract instructions from chat_options in kwargs
- Add unit tests to verify system_instructions are captured correctly

When using ChatClient with ChatOptions.instructions, the OpenTelemetry
LLM span was missing system messages in gen_ai.input.messages and the
gen_ai.system_instructions attribute was not being set.

This fix aligns the ChatClient-level tracing with the Agent-level
tracing which already correctly passes system_instructions.

Fixes #3163

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

* Add edge case tests for system_instructions

- Add test for empty string instructions (should not set attribute)
- Add test for list-type instructions (verify multiple items captured)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

* Simplify: use options.get('instructions') directly instead of kwargs.get('chat_options')

Addresses reviewer feedback:
- Removed unnecessary chat_options variable from kwargs
- Directly access instructions from the options parameter
- Updated tests to use dict syntax for options (TypedDict convention)

---------

Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
2026-01-20 14:04:31 +00:00
westeyandGitHub 3ec881509c .NET: Delete sync extension methods for agent (#3291)
* Delete sync extension methods for agent

* Fix comments and obsolete attribute

* Remove more sync methods.

* Fix naming and comments.

* Fix unit tests
2026-01-20 11:24:58 +00:00
dependabot[bot]GitHubdependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
8ee379d344 Bump tar from 7.4.3 to 7.5.3 in /python/packages/devui/frontend (#3267)
Bumps [tar](https://github.com/isaacs/node-tar) from 7.4.3 to 7.5.3.
- [Release notes](https://github.com/isaacs/node-tar/releases)
- [Changelog](https://github.com/isaacs/node-tar/blob/main/CHANGELOG.md)
- [Commits](https://github.com/isaacs/node-tar/compare/v7.4.3...v7.5.3)

---
updated-dependencies:
- dependency-name: tar
  dependency-version: 7.5.3
  dependency-type: indirect
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2026-01-20 07:26:34 +00:00
dependabot[bot]GitHubdependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2a43caefaa Bump ruff from 0.14.11 to 0.14.13 in /python (#3287)
Bumps [ruff](https://github.com/astral-sh/ruff) from 0.14.11 to 0.14.13.
- [Release notes](https://github.com/astral-sh/ruff/releases)
- [Changelog](https://github.com/astral-sh/ruff/blob/main/CHANGELOG.md)
- [Commits](https://github.com/astral-sh/ruff/compare/0.14.11...0.14.13)

---
updated-dependencies:
- dependency-name: ruff
  dependency-version: 0.14.13
  dependency-type: direct:development
  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>
2026-01-20 07:26:06 +00:00
dependabot[bot]GitHubdependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
f54248b79f Bump uv from 0.9.25 to 0.9.26 in /python (#3288)
Bumps [uv](https://github.com/astral-sh/uv) from 0.9.25 to 0.9.26.
- [Release notes](https://github.com/astral-sh/uv/releases)
- [Changelog](https://github.com/astral-sh/uv/blob/main/CHANGELOG.md)
- [Commits](https://github.com/astral-sh/uv/compare/0.9.25...0.9.26)

---
updated-dependencies:
- dependency-name: uv
  dependency-version: 0.9.26
  dependency-type: direct:development
  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>
2026-01-20 07:25:45 +00:00
Victor DibiaandGitHub 0f29637b86 fix #3171, ensure proper form rendering for int (#3201) 2026-01-20 07:25:12 +00:00
Evan MattsonandGitHub e0b9be7e08 Python: fix(declarative): Fix MCP tool connection not passed from YAML to Azure AI agent creation API (#3248)
* fix(declarative): Fix MCP tool connection not passed from YAML

* Add samples to README

* Fix mypy

* Fix mypy again

* Address PR comments
2026-01-20 07:24:20 +00:00
Dmytro StrukandGitHub 83e8965c8e Python: [BREAKING] Make response_format validation errors visible to users (#3274)
* Make response_format validation errors visible to users

* Small fix

* Addressed comments
2026-01-19 17:20:46 +00:00
Mark WallaceandGitHub 3c1be2a713 Update ignored checks in merge-gatekeeper workflow 2026-01-19 16:17:24 +00:00
dependabot[bot]GitHubdependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
467d3a60ed Bump actions/setup-dotnet from 5.0.1 to 5.1.0 (#3273)
Bumps [actions/setup-dotnet](https://github.com/actions/setup-dotnet) from 5.0.1 to 5.1.0.
- [Release notes](https://github.com/actions/setup-dotnet/releases)
- [Commits](https://github.com/actions/setup-dotnet/compare/v5.0.1...v5.1.0)

---
updated-dependencies:
- dependency-name: actions/setup-dotnet
  dependency-version: 5.1.0
  dependency-type: direct:production
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2026-01-19 15:58:02 +00:00
3243652df6 Python: Filter conversation_id when passing kwargs to agent as tool (#3266)
* Filter conversation_id when passing kwargs to agent as tool

* Small fix

* Update python/samples/getting_started/agents/azure_ai/README.md

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Update python/samples/getting_started/agents/openai/openai_responses_client_with_agent_as_tool.py

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Update python/samples/getting_started/agents/azure_ai/azure_ai_with_agent_as_tool.py

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
2026-01-19 12:54:06 +00:00
Dmytro StrukandGitHub 915df3b404 Python: Added rai_config to Azure AI agent creation (#3265)
* Add kwargs to create_agent method

* Added test for kwargs

* Addressed comment

* Added doc string
2026-01-19 12:51:27 +00:00
Dmytro StrukandGitHub f87e55ba33 Python: Fixed use_agent_middleware calling private _normalize_messages (#3264)
* Fix use_agent_middleware calling private _normalize_messages

* Fixed A2A and Copilot Studio agent
2026-01-19 12:50:14 +00:00
Dmytro StrukandGitHub 9bfa1a913c Python: Fixed Azure chat client for asynchronous filtering (#3260)
* Fixed Azure chat client for asynchronous filtering

* Updated test
2026-01-19 12:49:23 +00:00
Giles OdigweandGitHub 9e3b2fa09a Python: Update package version (#3258)
* package version 260116

* removed name tags
2026-01-16 21:06:30 +00:00
Dmytro StrukandGitHub 5687e13221 Python: [BREAKING] Renamed create_agent to as_agent (#3249)
* Renamed create_agent to as_agent

* Override for as_agent

* Added override
2026-01-16 19:21:52 +00:00
eoindoherty1andGitHub a151f10cc2 .NET Purview Middleware: Improve Background Job Runner Injection (#3256)
* Clean up background job dependency injection

* Fix xml documentation grammar
2026-01-16 19:20:35 +00:00
Dmytro StrukandGitHub b773830e4b Create/Get Agent API - fixes and example improvements (#3246) 2026-01-16 04:36:34 +00:00
Hao LuoandGitHub 975884f32d Python: (AG-UI) Support service-managed thread on AG-UI (#3136)
* added service thread support

* set service_thread_id to only supplied_thread_id

* uses raw_representation to extract the conversation_id

* removed accidental edit

* updated test to use raw_representation

* resolves copilot review feedback

* revert back StubAgent, since not used

* removed relative module import

* removed hasattr check per PR feedback
2026-01-16 03:28:13 +00:00
Dmytro StrukandGitHub b5ca0c8eda Python: Create/Get Agent API for OpenAI Assistants (#3208)
* Added provider implementation

* Added example with response format

* Small improvements
2026-01-15 22:52:32 +00:00
dependabot[bot]GitHubdependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
dd3e2b6e53 Bump azure-core from 1.37.0 to 1.38.0 in /python (#3209)
Bumps [azure-core](https://github.com/Azure/azure-sdk-for-python) from 1.37.0 to 1.38.0.
- [Release notes](https://github.com/Azure/azure-sdk-for-python/releases)
- [Commits](https://github.com/Azure/azure-sdk-for-python/compare/azure-core_1.37.0...azure-core_1.38.0)

---
updated-dependencies:
- dependency-name: azure-core
  dependency-version: 1.38.0
  dependency-type: indirect
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2026-01-15 22:25:41 +00:00
Dmytro StrukandGitHub 48d124efbe Python: Create/Get Agent API for Azure V1 (#3192)
* Added provider implementation for Azure AI V1

* Small fixes

* Fixed OpenAPI example

* Fixed local MCP example

* Fixed hosted MCP example

* Fixed file search sample

* Small fixes

* Resolved comments

* Doc updates
2026-01-15 22:19:03 +00:00
Dmytro StrukandGitHub 6e9420f614 Updated DurableAIAgent and fixed integration tests (#3241) 2026-01-15 21:43:26 +00:00
Dmytro StrukandGitHub 2ab859dd94 .NET: [BREAKING] Renamed CreateAIAgent/GetAIAgent to AsAIAgent (#3222)
* Renamed chat client extension method

* Additional renaming

* Updated documentation

* Fixed tests

* Small fix

* Small fix
2026-01-15 16:01:15 +00:00
CopilotGitHubrogerbarretocopilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
e192af93a7 .NET: Update Google.GenAI to 0.11.0 and remove polyfill implementations (#3232)
* Initial plan

* Update Google.GenAI to 0.11.0 and remove polyfill files

Co-authored-by: rogerbarreto <19890735+rogerbarreto@users.noreply.github.com>

---------

Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: rogerbarreto <19890735+rogerbarreto@users.noreply.github.com>
2026-01-15 15:16:32 +00:00
westeyandGitHub 3dbdecedda .NET: Merge AgentRunOptions.AdditionalProperties into ChatOptions.AdditionalProperties (#3184)
* Merge AgentRunOptions.AdditionalProperties into ChatOptions.AdditionalProperties

* Fix namespace and typo.
2026-01-15 12:15:24 +00:00
Evan MattsonandGitHub 15d0c34d9f Python: Properly configure structured outputs based on new options dict (#3213)
* Properly configure structured outputs based on new options dict

* Fix mypy
2026-01-15 11:41:46 +09:00
Evan MattsonandGitHub 620da7a829 Python: fix(ag-ui): add MCP tool support for AG-UI approval flows (#3212)
* add MCP tool support for AG-UI approval flows

* use attribute in place of property
2026-01-15 02:34:11 +00:00
Evan MattsonandGitHub 80b25a782b fix(workflows): rename WorkflowOutputEvent.source_executor_id to executor_id for API consistency (#3166) 2026-01-15 02:11:25 +00:00
Evan MattsonandGitHub ffe2e787ba Python: fix(core): correct FunctionResultContent ordering in WorkflowAgent.merge_updates (#3168)
* fix(core): simplify FunctionResultContent ordering in WorkflowAgent.merge_updates

* improve comment

* Fix name
2026-01-15 01:54:44 +00:00
557 changed files with 17643 additions and 10340 deletions
@@ -12,7 +12,7 @@ runs:
docker rm -f dts-emulator
fi
echo "Starting Durable Task Scheduler Emulator"
docker run -d --name dts-emulator -p 8080:8080 -p 8082:8082 mcr.microsoft.com/dts/dts-emulator:latest
docker run -d --name dts-emulator -p 8080:8080 -p 8082:8082 -e DTS_USE_DYNAMIC_TASK_HUBS=true mcr.microsoft.com/dts/dts-emulator:latest
echo "Waiting for Durable Task Scheduler Emulator to be ready"
timeout 30 bash -c 'until curl --silent http://localhost:8080/healthz; do sleep 1; done'
echo "Durable Task Scheduler Emulator is ready"
+1 -1
View File
@@ -95,7 +95,7 @@ jobs:
echo "COSMOS_EMULATOR_AVAILABLE=true" >> $env:GITHUB_ENV
- name: Setup dotnet
uses: actions/setup-dotnet@v5.0.1
uses: actions/setup-dotnet@v5.1.0
with:
global-json-file: ${{ github.workspace }}/dotnet/global.json
- name: Build dotnet solutions
+1 -1
View File
@@ -29,4 +29,4 @@ jobs:
token: ${{ secrets.GITHUB_TOKEN }}
timeout: 3600
interval: 30
ignored: CodeQL
ignored: CodeQL,CodeQL analysis (csharp)
@@ -34,9 +34,16 @@ jobs:
# because the workflow_run event does not have access to the PR number
# The PR number is needed to post the comment on the PR
run: |
PR_NUMBER=$(cat pr_number)
echo "PR number: $PR_NUMBER"
echo "PR_NUMBER=$PR_NUMBER" >> $GITHUB_ENV
if [ ! -s pr_number ]; then
echo "PR number file 'pr_number' is missing or empty"
exit 1
fi
PR_NUMBER=$(head -1 pr_number | tr -dc '0-9')
if [ -z "$PR_NUMBER" ]; then
echo "PR number file 'pr_number' does not contain a valid PR number"
exit 1
fi
echo "PR_NUMBER=$PR_NUMBER" >> "$GITHUB_ENV"
- name: Pytest coverage comment
id: coverageComment
uses: MishaKav/pytest-coverage-comment@v1.2.0
+1
View File
@@ -206,6 +206,7 @@ agents.md
WARP.md
**/memory-bank/
**/projectBrief.md
**/tmpclaude*
# Azurite storage emulator files
*/__azurite_db_blob__.json
+10 -10
View File
@@ -26,25 +26,25 @@
<PackageVersion Include="Azure.Identity" Version="1.17.1" />
<PackageVersion Include="Azure.Monitor.OpenTelemetry.Exporter" Version="1.4.0" />
<!-- Google Gemini -->
<PackageVersion Include="Google.GenAI" Version="0.9.0" />
<PackageVersion Include="Google.GenAI" Version="0.11.0" />
<PackageVersion Include="Mscc.GenerativeAI.Microsoft" Version="2.9.3" />
<!-- Microsoft.Azure.* -->
<PackageVersion Include="Microsoft.Azure.Cosmos" Version="3.54.0" />
<!-- Newtonsoft.Json -->
<PackageVersion Include="Newtonsoft.Json" Version="13.0.4" />
<!-- System.* -->
<PackageVersion Include="Microsoft.Bcl.AsyncInterfaces" Version="10.0.1" />
<PackageVersion Include="Microsoft.Bcl.AsyncInterfaces" Version="10.0.2" />
<PackageVersion Include="Microsoft.Bcl.HashCode" Version="6.0.0" />
<PackageVersion Include="System.ClientModel" Version="1.8.1" />
<PackageVersion Include="System.CodeDom" Version="10.0.0" />
<PackageVersion Include="System.Collections.Immutable" Version="10.0.0" />
<PackageVersion Include="System.CommandLine" Version="2.0.0-rc.2.25502.107" />
<PackageVersion Include="System.Diagnostics.DiagnosticSource" Version="10.0.1" />
<PackageVersion Include="System.Diagnostics.DiagnosticSource" Version="10.0.2" />
<PackageVersion Include="System.Linq.AsyncEnumerable" Version="10.0.0" />
<PackageVersion Include="System.Net.Http.Json" Version="10.0.0" />
<PackageVersion Include="System.Net.ServerSentEvents" Version="10.0.0" />
<PackageVersion Include="System.Text.Json" Version="10.0.1" />
<PackageVersion Include="System.Threading.Channels" Version="10.0.1" />
<PackageVersion Include="System.Text.Json" Version="10.0.2" />
<PackageVersion Include="System.Threading.Channels" Version="10.0.2" />
<PackageVersion Include="System.Threading.Tasks.Extensions" Version="4.6.3" />
<PackageVersion Include="System.Net.Security" Version="4.3.2" />
<!-- OpenTelemetry -->
@@ -61,9 +61,9 @@
<PackageVersion Include="Microsoft.AspNetCore.OpenApi" Version="10.0.0" />
<PackageVersion Include="Swashbuckle.AspNetCore.SwaggerUI" Version="10.0.0" />
<!-- Microsoft.Extensions.* -->
<PackageVersion Include="Microsoft.Extensions.AI" Version="10.1.1" />
<PackageVersion Include="Microsoft.Extensions.AI.Abstractions" Version="10.1.1" />
<PackageVersion Include="Microsoft.Extensions.AI.OpenAI" Version="10.1.1-preview.1.25612.2" />
<PackageVersion Include="Microsoft.Extensions.AI" Version="10.2.0" />
<PackageVersion Include="Microsoft.Extensions.AI.Abstractions" Version="10.2.0" />
<PackageVersion Include="Microsoft.Extensions.AI.OpenAI" Version="10.2.0-preview.1.26063.2" />
<PackageVersion Include="Microsoft.Extensions.Caching.Memory" Version="10.0.0" />
<PackageVersion Include="Microsoft.Extensions.Configuration" Version="10.0.0" />
<PackageVersion Include="Microsoft.Extensions.Configuration.Binder" Version="10.0.0" />
@@ -71,11 +71,11 @@
<PackageVersion Include="Microsoft.Extensions.Configuration.Json" Version="10.0.0" />
<PackageVersion Include="Microsoft.Extensions.Configuration.UserSecrets" Version="10.0.0" />
<PackageVersion Include="Microsoft.Extensions.DependencyInjection" Version="10.0.0" />
<PackageVersion Include="Microsoft.Extensions.DependencyInjection.Abstractions" Version="10.0.1" />
<PackageVersion Include="Microsoft.Extensions.DependencyInjection.Abstractions" Version="10.0.2" />
<PackageVersion Include="Microsoft.Extensions.Hosting" Version="10.0.0" />
<PackageVersion Include="Microsoft.Extensions.Http.Resilience" Version="10.0.0" />
<PackageVersion Include="Microsoft.Extensions.Logging" Version="10.0.0" />
<PackageVersion Include="Microsoft.Extensions.Logging.Abstractions" Version="10.0.1" />
<PackageVersion Include="Microsoft.Extensions.Logging.Abstractions" Version="10.0.2" />
<PackageVersion Include="Microsoft.Extensions.Logging.Console" Version="10.0.0" />
<PackageVersion Include="Microsoft.Extensions.ServiceDiscovery" Version="10.0.0" />
<PackageVersion Include="Microsoft.Extensions.VectorData.Abstractions" Version="9.7.0" />
+1 -1
View File
@@ -21,7 +21,7 @@ var deploymentName = Environment.GetEnvironmentVariable("AZURE_OPENAI_DEPLOYMENT
var agent = new AzureOpenAIClient(new Uri(endpoint), new AzureCliCredential())
.GetOpenAIResponseClient(deploymentName)
.CreateAIAgent(name: "HaikuBot", instructions: "You are an upbeat assistant that writes beautifully.");
.AsAIAgent(name: "HaikuBot", instructions: "You are an upbeat assistant that writes beautifully.");
Console.WriteLine(await agent.RunAsync("Write a haiku about Microsoft Agent Framework."));
```
+13
View File
@@ -35,6 +35,18 @@
<Project Path="samples/AzureFunctions/07_AgentAsMcpTool/07_AgentAsMcpTool.csproj" />
<Project Path="samples/AzureFunctions/08_ReliableStreaming/08_ReliableStreaming.csproj" />
</Folder>
<Folder Name="/Samples/DurableAgents/">
<File Path="samples/DurableAgents/ConsoleApps/README.md" />
</Folder>
<Folder Name="/Samples/DurableAgents/ConsoleApps/">
<Project Path="samples/DurableAgents/ConsoleApps/01_SingleAgent/01_SingleAgent.csproj" />
<Project Path="samples/DurableAgents/ConsoleApps/02_AgentOrchestration_Chaining/02_AgentOrchestration_Chaining.csproj" />
<Project Path="samples/DurableAgents/ConsoleApps/03_AgentOrchestration_Concurrency/03_AgentOrchestration_Concurrency.csproj" />
<Project Path="samples/DurableAgents/ConsoleApps/04_AgentOrchestration_Conditionals/04_AgentOrchestration_Conditionals.csproj" />
<Project Path="samples/DurableAgents/ConsoleApps/05_AgentOrchestration_HITL/05_AgentOrchestration_HITL.csproj" />
<Project Path="samples/DurableAgents/ConsoleApps/06_LongRunningTools/06_LongRunningTools.csproj" />
<Project Path="samples/DurableAgents/ConsoleApps/07_ReliableStreaming/07_ReliableStreaming.csproj" />
</Folder>
<Folder Name="/Samples/GettingStarted/">
<File Path="samples/GettingStarted/README.md" />
</Folder>
@@ -81,6 +93,7 @@
<Project Path="samples/GettingStarted/Agents/Agent_Step17_BackgroundResponses/Agent_Step17_BackgroundResponses.csproj" />
<Project Path="samples/GettingStarted/Agents/Agent_Step18_DeepResearch/Agent_Step18_DeepResearch.csproj" />
<Project Path="samples/GettingStarted/Agents/Agent_Step19_Declarative/Agent_Step19_Declarative.csproj" />
<Project Path="samples/GettingStarted/Agents/Agent_Step20_AdditionalAIContext/Agent_Step20_AdditionalAIContext.csproj" />
</Folder>
<Folder Name="/Samples/GettingStarted/DeclarativeAgents/">
<Project Path="samples/GettingStarted/DeclarativeAgents/ChatClient/DeclarativeChatClientAgents.csproj" />
+3 -3
View File
@@ -2,9 +2,9 @@
<PropertyGroup>
<!-- Central version prefix - applies to all nuget packages. -->
<VersionPrefix>1.0.0</VersionPrefix>
<PackageVersion Condition="'$(VersionSuffix)' != ''">$(VersionPrefix)-$(VersionSuffix).260108.1</PackageVersion>
<PackageVersion Condition="'$(VersionSuffix)' == ''">$(VersionPrefix)-preview.260108.1</PackageVersion>
<GitTag>1.0.0-preview.260108.1</GitTag>
<PackageVersion Condition="'$(VersionSuffix)' != ''">$(VersionPrefix)-$(VersionSuffix).260121.1</PackageVersion>
<PackageVersion Condition="'$(VersionSuffix)' == ''">$(VersionPrefix)-preview.260121.1</PackageVersion>
<GitTag>1.0.0-preview.260121.1</GitTag>
<Configurations>Debug;Release;Publish</Configurations>
<IsPackable>true</IsPackable>
@@ -29,7 +29,7 @@ internal sealed class HostClientAgent
// Create the agent that uses the remote agents as tools
this.Agent = new OpenAIClient(new ApiKeyCredential(apiKey))
.GetChatClient(modelId)
.CreateAIAgent(instructions: "You specialize in handling queries for users and using your tools to provide answers.", name: "HostClient", tools: tools);
.AsAIAgent(instructions: "You specialize in handling queries for users and using your tools to provide answers.", name: "HostClient", tools: tools);
}
catch (Exception ex)
{
@@ -35,7 +35,7 @@ internal static class HostAgentFactory
{
AIAgent agent = new OpenAIClient(apiKey)
.GetChatClient(model)
.CreateAIAgent(instructions, name, tools: tools);
.AsAIAgent(instructions, name, tools: tools);
AgentCard agentCard = agentType.ToUpperInvariant() switch
{
@@ -83,7 +83,7 @@ public static class Program
serverUrl,
jsonSerializerOptions: AGUIClientSerializerContext.Default.Options);
AIAgent agent = chatClient.CreateAIAgent(
AIAgent agent = chatClient.AsAIAgent(
name: "agui-client",
description: "AG-UI Client Agent",
tools: [changeBackground, readClientClimateSensors]);
@@ -33,7 +33,7 @@ internal static class ChatClientAgentFactory
{
ChatClient chatClient = s_azureOpenAIClient!.GetChatClient(s_deploymentName!);
return chatClient.AsIChatClient().CreateAIAgent(
return chatClient.AsIChatClient().AsAIAgent(
name: "AgenticChat",
description: "A simple chat agent using Azure OpenAI");
}
@@ -42,7 +42,7 @@ internal static class ChatClientAgentFactory
{
ChatClient chatClient = s_azureOpenAIClient!.GetChatClient(s_deploymentName!);
return chatClient.AsIChatClient().CreateAIAgent(
return chatClient.AsIChatClient().AsAIAgent(
name: "BackendToolRenderer",
description: "An agent that can render backend tools using Azure OpenAI",
tools: [AIFunctionFactory.Create(
@@ -56,7 +56,7 @@ internal static class ChatClientAgentFactory
{
ChatClient chatClient = s_azureOpenAIClient!.GetChatClient(s_deploymentName!);
return chatClient.AsIChatClient().CreateAIAgent(
return chatClient.AsIChatClient().AsAIAgent(
name: "HumanInTheLoopAgent",
description: "An agent that involves human feedback in its decision-making process using Azure OpenAI");
}
@@ -65,7 +65,7 @@ internal static class ChatClientAgentFactory
{
ChatClient chatClient = s_azureOpenAIClient!.GetChatClient(s_deploymentName!);
return chatClient.AsIChatClient().CreateAIAgent(
return chatClient.AsIChatClient().AsAIAgent(
name: "ToolBasedGenerativeUIAgent",
description: "An agent that uses tools to generate user interfaces using Azure OpenAI");
}
@@ -73,7 +73,7 @@ internal static class ChatClientAgentFactory
public static AIAgent CreateAgenticUI(JsonSerializerOptions options)
{
ChatClient chatClient = s_azureOpenAIClient!.GetChatClient(s_deploymentName!);
var baseAgent = chatClient.AsIChatClient().CreateAIAgent(new ChatClientAgentOptions
var baseAgent = chatClient.AsIChatClient().AsAIAgent(new ChatClientAgentOptions
{
Name = "AgenticUIAgent",
Description = "An agent that generates agentic user interfaces using Azure OpenAI",
@@ -116,7 +116,7 @@ internal static class ChatClientAgentFactory
{
ChatClient chatClient = s_azureOpenAIClient!.GetChatClient(s_deploymentName!);
var baseAgent = chatClient.AsIChatClient().CreateAIAgent(
var baseAgent = chatClient.AsIChatClient().AsAIAgent(
name: "SharedStateAgent",
description: "An agent that demonstrates shared state patterns using Azure OpenAI");
@@ -127,7 +127,7 @@ internal static class ChatClientAgentFactory
{
ChatClient chatClient = s_azureOpenAIClient!.GetChatClient(s_deploymentName!);
var baseAgent = chatClient.AsIChatClient().CreateAIAgent(new ChatClientAgentOptions
var baseAgent = chatClient.AsIChatClient().AsAIAgent(new ChatClientAgentOptions
{
Name = "PredictiveStateUpdatesAgent",
Description = "An agent that demonstrates predictive state updates using Azure OpenAI",
@@ -23,7 +23,7 @@ var agent = new AzureOpenAIClient(
new Uri(endpoint),
new DefaultAzureCredential())
.GetChatClient(deploymentName)
.CreateAIAgent(
.AsAIAgent(
name: "AGUIAssistant",
tools: [
AIFunctionFactory.Create(
+2 -2
View File
@@ -119,7 +119,7 @@ The `AGUIServer` uses the `MapAGUI` extension method to expose an agent through
```csharp
AIAgent agent = new OpenAIClient(apiKey)
.GetChatClient(model)
.CreateAIAgent(
.AsAIAgent(
instructions: "You are a helpful assistant.",
name: "AGUIAssistant");
@@ -144,7 +144,7 @@ var chatClient = new AGUIChatClient(
modelId: "agui-client",
jsonSerializerOptions: null);
AIAgent agent = chatClient.CreateAIAgent(
AIAgent agent = chatClient.AsAIAgent(
instructions: null,
name: "agui-client",
description: "AG-UI Client Agent",
+2 -2
View File
@@ -74,7 +74,7 @@ AzureOpenAIClient azureOpenAIClient = new AzureOpenAIClient(
ChatClient chatClient = azureOpenAIClient.GetChatClient(deploymentName);
// Create AI agent
ChatClientAgent agent = chatClient.AsIChatClient().CreateAIAgent(
ChatClientAgent agent = chatClient.AsIChatClient().AsAIAgent(
name: "ChatAssistant",
instructions: "You are a helpful assistant.");
@@ -162,7 +162,7 @@ dotnet run
Edit the instructions in `Server/Program.cs`:
```csharp
ChatClientAgent agent = chatClient.AsIChatClient().CreateAIAgent(
ChatClientAgent agent = chatClient.AsIChatClient().AsAIAgent(
name: "ChatAssistant",
instructions: "You are a helpful coding assistant specializing in C# and .NET.");
```
+1 -1
View File
@@ -25,7 +25,7 @@ AzureOpenAIClient azureOpenAIClient = new(
ChatClient chatClient = azureOpenAIClient.GetChatClient(deploymentName);
ChatClientAgent agent = chatClient.AsIChatClient().CreateAIAgent(
ChatClientAgent agent = chatClient.AsIChatClient().AsAIAgent(
name: "ChatAssistant",
instructions: "You are a helpful assistant.");
@@ -25,7 +25,7 @@ AzureOpenAIClient client = !string.IsNullOrEmpty(azureOpenAiKey)
const string JokerName = "Joker";
const string JokerInstructions = "You are good at telling jokes.";
AIAgent agent = client.GetChatClient(deploymentName).CreateAIAgent(JokerInstructions, JokerName);
AIAgent agent = client.GetChatClient(deploymentName).AsAIAgent(JokerInstructions, JokerName);
// Configure the function app to host the AI agent.
// This will automatically generate HTTP API endpoints for the agent.
@@ -29,7 +29,7 @@ const string WriterInstructions =
when given an improved sentence you polish it further.
""";
AIAgent writerAgent = client.GetChatClient(deploymentName).CreateAIAgent(WriterInstructions, WriterName);
AIAgent writerAgent = client.GetChatClient(deploymentName).AsAIAgent(WriterInstructions, WriterName);
using IHost app = FunctionsApplication
.CreateBuilder(args)
@@ -28,8 +28,8 @@ const string PhysicistInstructions = "You are an expert in physics. You answer q
const string ChemistName = "ChemistAgent";
const string ChemistInstructions = "You are an expert in chemistry. You answer questions from a chemistry perspective.";
AIAgent physicistAgent = client.GetChatClient(deploymentName).CreateAIAgent(PhysicistInstructions, PhysicistName);
AIAgent chemistAgent = client.GetChatClient(deploymentName).CreateAIAgent(ChemistInstructions, ChemistName);
AIAgent physicistAgent = client.GetChatClient(deploymentName).AsAIAgent(PhysicistInstructions, PhysicistName);
AIAgent chemistAgent = client.GetChatClient(deploymentName).AsAIAgent(ChemistInstructions, ChemistName);
using IHost app = FunctionsApplication
.CreateBuilder(args)
@@ -29,10 +29,10 @@ const string EmailAssistantName = "EmailAssistantAgent";
const string EmailAssistantInstructions = "You are an email assistant that helps users draft responses to emails with professionalism.";
AIAgent spamDetectionAgent = client.GetChatClient(deploymentName)
.CreateAIAgent(SpamDetectionInstructions, SpamDetectionName);
.AsAIAgent(SpamDetectionInstructions, SpamDetectionName);
AIAgent emailAssistantAgent = client.GetChatClient(deploymentName)
.CreateAIAgent(EmailAssistantInstructions, EmailAssistantName);
.AsAIAgent(EmailAssistantInstructions, EmailAssistantName);
using IHost app = FunctionsApplication
.CreateBuilder(args)
@@ -29,7 +29,7 @@ const string WriterInstructions =
You write engaging, informative, and well-structured content that follows best practices for readability and accuracy.
""";
AIAgent writerAgent = client.GetChatClient(deploymentName).CreateAIAgent(WriterInstructions, WriterName);
AIAgent writerAgent = client.GetChatClient(deploymentName).AsAIAgent(WriterInstructions, WriterName);
using IHost app = FunctionsApplication
.CreateBuilder(args)
@@ -33,7 +33,7 @@ const string WriterAgentInstructions =
You write engaging, informative, and well-structured content that follows best practices for readability and accuracy.
""";
AIAgent writerAgent = client.GetChatClient(deploymentName).CreateAIAgent(WriterAgentInstructions, WriterAgentName);
AIAgent writerAgent = client.GetChatClient(deploymentName).AsAIAgent(WriterAgentInstructions, WriterAgentName);
// Agent that can start content generation workflows using tools
const string PublisherAgentName = "Publisher";
@@ -57,7 +57,7 @@ using IHost app = FunctionsApplication
// Initialize the tools to be used by the agent.
Tools publisherTools = new(sp.GetRequiredService<ILogger<Tools>>());
return client.GetChatClient(deploymentName).CreateAIAgent(
return client.GetChatClient(deploymentName).AsAIAgent(
instructions: PublisherAgentInstructions,
name: PublisherAgentName,
services: sp,
@@ -28,13 +28,13 @@ AzureOpenAIClient client = !string.IsNullOrEmpty(azureOpenAiKey)
: new AzureOpenAIClient(new Uri(endpoint), new AzureCliCredential());
// Define three AI agents we are going to use in this application.
AIAgent agent1 = client.GetChatClient(deploymentName).CreateAIAgent("You are good at telling jokes.", "Joker");
AIAgent agent1 = client.GetChatClient(deploymentName).AsAIAgent("You are good at telling jokes.", "Joker");
AIAgent agent2 = client.GetChatClient(deploymentName)
.CreateAIAgent("Check stock prices.", "StockAdvisor");
.AsAIAgent("Check stock prices.", "StockAdvisor");
AIAgent agent3 = client.GetChatClient(deploymentName)
.CreateAIAgent("Recommend plants.", "PlantAdvisor", description: "Get plant recommendations.");
.AsAIAgent("Recommend plants.", "PlantAdvisor", description: "Get plant recommendations.");
using IHost app = FunctionsApplication
.CreateBuilder(args)
@@ -70,7 +70,7 @@ FunctionsApplicationBuilder builder = FunctionsApplication
// Define the Travel Planner agent with tools for weather and events
options.AddAIAgentFactory(TravelPlannerName, sp =>
{
return client.GetChatClient(deploymentName).CreateAIAgent(
return client.GetChatClient(deploymentName).AsAIAgent(
instructions: TravelPlannerInstructions,
name: TravelPlannerName,
services: sp,
@@ -0,0 +1,30 @@
<Project Sdk="Microsoft.NET.Sdk">
<PropertyGroup>
<TargetFrameworks>net10.0</TargetFrameworks>
<OutputType>Exe</OutputType>
<ImplicitUsings>enable</ImplicitUsings>
<Nullable>enable</Nullable>
<AssemblyName>SingleAgent</AssemblyName>
<RootNamespace>SingleAgent</RootNamespace>
</PropertyGroup>
<ItemGroup>
<PackageReference Include="Azure.AI.OpenAI" />
<PackageReference Include="Azure.Identity" />
<PackageReference Include="Microsoft.DurableTask.Client.AzureManaged" />
<PackageReference Include="Microsoft.DurableTask.Worker.AzureManaged" />
<PackageReference Include="Microsoft.Extensions.Hosting" />
</ItemGroup>
<!-- Local projects that should be switched to package references when using the sample outside of this MAF repo -->
<!--
<ItemGroup>
<PackageReference Include="Microsoft.Agents.AI.DurableTask" />
<PackageReference Include="Microsoft.Agents.AI.OpenAI" />
</ItemGroup>
-->
<ItemGroup>
<ProjectReference Include="..\..\..\..\src\Microsoft.Agents.AI.DurableTask\Microsoft.Agents.AI.DurableTask.csproj" />
<ProjectReference Include="..\..\..\..\src\Microsoft.Agents.AI.OpenAI\Microsoft.Agents.AI.OpenAI.csproj" />
</ItemGroup>
</Project>
@@ -0,0 +1,103 @@
// Copyright (c) Microsoft. All rights reserved.
using Azure;
using Azure.AI.OpenAI;
using Azure.Identity;
using Microsoft.Agents.AI;
using Microsoft.Agents.AI.DurableTask;
using Microsoft.DurableTask.Client.AzureManaged;
using Microsoft.DurableTask.Worker.AzureManaged;
using Microsoft.Extensions.AI;
using Microsoft.Extensions.DependencyInjection;
using Microsoft.Extensions.Hosting;
using Microsoft.Extensions.Logging;
using OpenAI.Chat;
// Get the Azure OpenAI endpoint and deployment name from environment variables.
string endpoint = Environment.GetEnvironmentVariable("AZURE_OPENAI_ENDPOINT")
?? throw new InvalidOperationException("AZURE_OPENAI_ENDPOINT is not set.");
string deploymentName = Environment.GetEnvironmentVariable("AZURE_OPENAI_DEPLOYMENT")
?? throw new InvalidOperationException("AZURE_OPENAI_DEPLOYMENT is not set.");
// Get DTS connection string from environment variable
string dtsConnectionString = Environment.GetEnvironmentVariable("DURABLE_TASK_SCHEDULER_CONNECTION_STRING")
?? "Endpoint=http://localhost:8080;TaskHub=default;Authentication=None";
// Use Azure Key Credential if provided, otherwise use Azure CLI Credential.
string? azureOpenAiKey = Environment.GetEnvironmentVariable("AZURE_OPENAI_KEY");
AzureOpenAIClient client = !string.IsNullOrEmpty(azureOpenAiKey)
? new AzureOpenAIClient(new Uri(endpoint), new AzureKeyCredential(azureOpenAiKey))
: new AzureOpenAIClient(new Uri(endpoint), new AzureCliCredential());
// Set up an AI agent following the standard Microsoft Agent Framework pattern.
const string JokerName = "Joker";
const string JokerInstructions = "You are good at telling jokes.";
AIAgent agent = client.GetChatClient(deploymentName).AsAIAgent(JokerInstructions, JokerName);
// Configure the console app to host the AI agent.
IHost host = Host.CreateDefaultBuilder(args)
.ConfigureLogging(logging => logging.SetMinimumLevel(LogLevel.Warning))
.ConfigureServices(services =>
{
services.ConfigureDurableAgents(
options => options.AddAIAgent(agent, timeToLive: TimeSpan.FromHours(1)),
workerBuilder: builder => builder.UseDurableTaskScheduler(dtsConnectionString),
clientBuilder: builder => builder.UseDurableTaskScheduler(dtsConnectionString));
})
.Build();
await host.StartAsync();
// Get the agent proxy from services
IServiceProvider services = host.Services;
AIAgent agentProxy = services.GetRequiredKeyedService<AIAgent>(JokerName);
// Console colors for better UX
Console.ForegroundColor = ConsoleColor.Cyan;
Console.WriteLine("=== Single Agent Console Sample ===");
Console.ResetColor();
Console.WriteLine("Enter a message for the Joker agent (or 'exit' to quit):");
Console.WriteLine();
// Create a thread for the conversation
AgentThread thread = await agentProxy.GetNewThreadAsync();
while (true)
{
// Read input from stdin
Console.ForegroundColor = ConsoleColor.Yellow;
Console.Write("You: ");
Console.ResetColor();
string? input = Console.ReadLine();
if (string.IsNullOrWhiteSpace(input) || input.Equals("exit", StringComparison.OrdinalIgnoreCase))
{
break;
}
// Run the agent
Console.ForegroundColor = ConsoleColor.Green;
Console.Write("Joker: ");
Console.ResetColor();
try
{
AgentResponse agentResponse = await agentProxy.RunAsync(
message: input,
thread: thread,
cancellationToken: CancellationToken.None);
Console.WriteLine(agentResponse.Text);
Console.WriteLine();
}
catch (Exception ex)
{
Console.ForegroundColor = ConsoleColor.Red;
Console.Error.WriteLine($"Error: {ex.Message}");
Console.ResetColor();
Console.WriteLine();
}
}
await host.StopAsync();
@@ -0,0 +1,56 @@
# Single Agent Sample
This sample demonstrates how to use the durable agents extension to create a simple console app that hosts a single AI agent and provides interactive conversation via stdin/stdout.
## Key Concepts Demonstrated
- Using the Microsoft Agent Framework to define a simple AI agent with a name and instructions.
- Registering durable agents with the console app and running them interactively.
- Conversation management (via threads) for isolated interactions.
## Environment Setup
See the [README.md](../README.md) file in the parent directory for more information on how to configure the environment, including how to install and run common sample dependencies.
## Running the Sample
With the environment setup, you can run the sample:
```bash
cd dotnet/samples/DurableAgents/ConsoleApps/01_SingleAgent
dotnet run --framework net10.0
```
The app will prompt you for input. You can interact with the Joker agent:
```text
=== Single Agent Console Sample ===
Enter a message for the Joker agent (or 'exit' to quit):
You: Tell me a joke about a pirate.
Joker: Why don't pirates ever learn the alphabet? Because they always get stuck at "C"!
You: Now explain the joke.
Joker: The joke plays on the word "sea" (C), which pirates are famously associated with...
You: exit
```
## Scriptable Usage
You can also pipe input to the app for scriptable usage:
```bash
echo "Tell me a joke about a pirate." | dotnet run
```
The app will read from stdin, process the input, and write the response to stdout.
## Viewing Agent State
You can view the state of the agent in the Durable Task Scheduler dashboard:
1. Open your browser and navigate to `http://localhost:8082`
2. In the dashboard, you can view the state of the Joker agent, including its conversation history and current state
The agent maintains conversation state across multiple interactions, and you can inspect this state in the dashboard to understand how the durable agents extension manages conversation context.
@@ -0,0 +1,30 @@
<Project Sdk="Microsoft.NET.Sdk">
<PropertyGroup>
<TargetFrameworks>net10.0</TargetFrameworks>
<OutputType>Exe</OutputType>
<ImplicitUsings>enable</ImplicitUsings>
<Nullable>enable</Nullable>
<AssemblyName>AgentOrchestration_Chaining</AssemblyName>
<RootNamespace>AgentOrchestration_Chaining</RootNamespace>
</PropertyGroup>
<ItemGroup>
<PackageReference Include="Azure.AI.OpenAI" />
<PackageReference Include="Azure.Identity" />
<PackageReference Include="Microsoft.DurableTask.Client.AzureManaged" />
<PackageReference Include="Microsoft.DurableTask.Worker.AzureManaged" />
<PackageReference Include="Microsoft.Extensions.Hosting" />
</ItemGroup>
<!-- Local projects that should be switched to package references when using the sample outside of this MAF repo -->
<!--
<ItemGroup>
<PackageReference Include="Microsoft.Agents.AI.DurableTask" />
<PackageReference Include="Microsoft.Agents.AI.OpenAI" />
</ItemGroup>
-->
<ItemGroup>
<ProjectReference Include="..\..\..\..\src\Microsoft.Agents.AI.DurableTask\Microsoft.Agents.AI.DurableTask.csproj" />
<ProjectReference Include="..\..\..\..\src\Microsoft.Agents.AI.OpenAI\Microsoft.Agents.AI.OpenAI.csproj" />
</ItemGroup>
</Project>
@@ -0,0 +1,6 @@
// Copyright (c) Microsoft. All rights reserved.
namespace AgentOrchestration_Chaining;
// Response model
public sealed record TextResponse(string Text);
@@ -0,0 +1,148 @@
// Copyright (c) Microsoft. All rights reserved.
using AgentOrchestration_Chaining;
using Azure;
using Azure.AI.OpenAI;
using Azure.Identity;
using Microsoft.Agents.AI;
using Microsoft.Agents.AI.DurableTask;
using Microsoft.DurableTask;
using Microsoft.DurableTask.Client;
using Microsoft.DurableTask.Client.AzureManaged;
using Microsoft.DurableTask.Worker;
using Microsoft.DurableTask.Worker.AzureManaged;
using Microsoft.Extensions.DependencyInjection;
using Microsoft.Extensions.Hosting;
using Microsoft.Extensions.Logging;
using OpenAI.Chat;
using Environment = System.Environment;
// Get the Azure OpenAI endpoint and deployment name from environment variables.
string endpoint = Environment.GetEnvironmentVariable("AZURE_OPENAI_ENDPOINT")
?? throw new InvalidOperationException("AZURE_OPENAI_ENDPOINT is not set.");
string deploymentName = Environment.GetEnvironmentVariable("AZURE_OPENAI_DEPLOYMENT")
?? throw new InvalidOperationException("AZURE_OPENAI_DEPLOYMENT is not set.");
// Get DTS connection string from environment variable
string dtsConnectionString = Environment.GetEnvironmentVariable("DURABLE_TASK_SCHEDULER_CONNECTION_STRING")
?? "Endpoint=http://localhost:8080;TaskHub=default;Authentication=None";
// Use Azure Key Credential if provided, otherwise use Azure CLI Credential.
string? azureOpenAiKey = Environment.GetEnvironmentVariable("AZURE_OPENAI_KEY");
AzureOpenAIClient client = !string.IsNullOrEmpty(azureOpenAiKey)
? new AzureOpenAIClient(new Uri(endpoint), new AzureKeyCredential(azureOpenAiKey))
: new AzureOpenAIClient(new Uri(endpoint), new AzureCliCredential());
// Single agent used by the orchestration to demonstrate sequential calls on the same thread.
const string WriterName = "WriterAgent";
const string WriterInstructions =
"""
You refine short pieces of text. When given an initial sentence you enhance it;
when given an improved sentence you polish it further.
""";
AIAgent writerAgent = client.GetChatClient(deploymentName).AsAIAgent(WriterInstructions, WriterName);
// Orchestrator function
static async Task<string> RunOrchestratorAsync(TaskOrchestrationContext context)
{
DurableAIAgent writer = context.GetAgent("WriterAgent");
AgentThread writerThread = await writer.GetNewThreadAsync();
AgentResponse<TextResponse> initial = await writer.RunAsync<TextResponse>(
message: "Write a concise inspirational sentence about learning.",
thread: writerThread);
AgentResponse<TextResponse> refined = await writer.RunAsync<TextResponse>(
message: $"Improve this further while keeping it under 25 words: {initial.Result.Text}",
thread: writerThread);
return refined.Result.Text;
}
// Configure the console app to host the AI agent.
IHost host = Host.CreateDefaultBuilder(args)
.ConfigureLogging(loggingBuilder => loggingBuilder.SetMinimumLevel(LogLevel.Warning))
.ConfigureServices(services =>
{
services.ConfigureDurableAgents(
options => options.AddAIAgent(writerAgent),
workerBuilder: builder =>
{
builder.UseDurableTaskScheduler(dtsConnectionString);
builder.AddTasks(registry => registry.AddOrchestratorFunc(nameof(RunOrchestratorAsync), RunOrchestratorAsync));
},
clientBuilder: builder => builder.UseDurableTaskScheduler(dtsConnectionString));
})
.Build();
await host.StartAsync();
DurableTaskClient durableClient = host.Services.GetRequiredService<DurableTaskClient>();
// Console colors for better UX
Console.ForegroundColor = ConsoleColor.Cyan;
Console.WriteLine("=== Single Agent Orchestration Chaining Sample ===");
Console.ResetColor();
Console.WriteLine("Starting orchestration...");
Console.WriteLine();
try
{
// Start the orchestration
string instanceId = await durableClient.ScheduleNewOrchestrationInstanceAsync(
orchestratorName: nameof(RunOrchestratorAsync));
Console.ForegroundColor = ConsoleColor.Gray;
Console.WriteLine($"Orchestration started with instance ID: {instanceId}");
Console.WriteLine("Waiting for completion...");
Console.ResetColor();
// Wait for orchestration to complete
OrchestrationMetadata status = await durableClient.WaitForInstanceCompletionAsync(
instanceId,
getInputsAndOutputs: true,
CancellationToken.None);
Console.WriteLine();
if (status.RuntimeStatus == OrchestrationRuntimeStatus.Completed)
{
Console.ForegroundColor = ConsoleColor.Green;
Console.WriteLine("✓ Orchestration completed successfully!");
Console.ResetColor();
Console.WriteLine();
Console.ForegroundColor = ConsoleColor.Yellow;
Console.Write("Result: ");
Console.ResetColor();
Console.WriteLine(status.ReadOutputAs<string>());
}
else if (status.RuntimeStatus == OrchestrationRuntimeStatus.Failed)
{
Console.ForegroundColor = ConsoleColor.Red;
Console.WriteLine("✗ Orchestration failed!");
Console.ResetColor();
if (status.FailureDetails != null)
{
Console.WriteLine($"Error: {status.FailureDetails.ErrorMessage}");
}
Environment.Exit(1);
}
else
{
Console.ForegroundColor = ConsoleColor.Yellow;
Console.WriteLine($"Orchestration status: {status.RuntimeStatus}");
Console.ResetColor();
}
}
catch (Exception ex)
{
Console.ForegroundColor = ConsoleColor.Red;
Console.Error.WriteLine($"Error: {ex.Message}");
Console.ResetColor();
Environment.Exit(1);
}
finally
{
await host.StopAsync();
}
@@ -0,0 +1,53 @@
# Single Agent Orchestration Sample
This sample demonstrates how to use the durable agents extension to create a simple console app that orchestrates sequential calls to a single AI agent using the same conversation thread for context continuity.
## Key Concepts Demonstrated
- Orchestrating multiple interactions with the same agent in a deterministic order
- Using the same `AgentThread` across multiple calls to maintain conversational context
- Durable orchestration with automatic checkpointing and resumption from failures
- Waiting for orchestration completion using `WaitForInstanceCompletionAsync`
## Environment Setup
See the [README.md](../README.md) file in the parent directory for more information on how to configure the environment, including how to install and run common sample dependencies.
## Running the Sample
With the environment setup, you can run the sample:
```bash
cd dotnet/samples/DurableAgents/ConsoleApps/02_AgentOrchestration_Chaining
dotnet run --framework net10.0
```
The app will start the orchestration, wait for it to complete, and display the result:
```text
=== Single Agent Orchestration Chaining Sample ===
Starting orchestration...
Orchestration started with instance ID: 86313f1d45fb42eeb50b1852626bf3ff
Waiting for completion...
✓ Orchestration completed successfully!
Result: Learning serves as the key, opening doors to boundless opportunities and a brighter future.
```
The orchestration will proceed to run the WriterAgent twice in sequence:
1. First, it writes an inspirational sentence about learning
2. Then, it refines the initial output using the same conversation thread
## Viewing Orchestration State
You can view the state of the orchestration in the Durable Task Scheduler dashboard:
1. Open your browser and navigate to `http://localhost:8082`
2. In the dashboard, you can see:
- **Orchestrations**: View the orchestration instance, including its runtime status, input, output, and execution history
- **Agents**: View the state of the WriterAgent, including conversation history maintained across the orchestration steps
The orchestration instance ID is displayed in the console output. You can use this ID to find the specific orchestration in the dashboard and inspect its execution details, including the sequence of agent calls and their results.
@@ -0,0 +1,30 @@
<Project Sdk="Microsoft.NET.Sdk">
<PropertyGroup>
<TargetFrameworks>net10.0</TargetFrameworks>
<OutputType>Exe</OutputType>
<ImplicitUsings>enable</ImplicitUsings>
<Nullable>enable</Nullable>
<AssemblyName>AgentOrchestration_Concurrency</AssemblyName>
<RootNamespace>AgentOrchestration_Concurrency</RootNamespace>
</PropertyGroup>
<ItemGroup>
<PackageReference Include="Azure.AI.OpenAI" />
<PackageReference Include="Azure.Identity" />
<PackageReference Include="Microsoft.DurableTask.Client.AzureManaged" />
<PackageReference Include="Microsoft.DurableTask.Worker.AzureManaged" />
<PackageReference Include="Microsoft.Extensions.Hosting" />
</ItemGroup>
<!-- Local projects that should be switched to package references when using the sample outside of this MAF repo -->
<!--
<ItemGroup>
<PackageReference Include="Microsoft.Agents.AI.DurableTask" />
<PackageReference Include="Microsoft.Agents.AI.OpenAI" />
</ItemGroup>
-->
<ItemGroup>
<ProjectReference Include="..\..\..\..\src\Microsoft.Agents.AI.DurableTask\Microsoft.Agents.AI.DurableTask.csproj" />
<ProjectReference Include="..\..\..\..\src\Microsoft.Agents.AI.OpenAI\Microsoft.Agents.AI.OpenAI.csproj" />
</ItemGroup>
</Project>
@@ -0,0 +1,6 @@
// Copyright (c) Microsoft. All rights reserved.
namespace AgentOrchestration_Concurrency;
// Response model
public sealed record TextResponse(string Text);
@@ -0,0 +1,191 @@
// Copyright (c) Microsoft. All rights reserved.
using System.Text.Json;
using AgentOrchestration_Concurrency;
using Azure;
using Azure.AI.OpenAI;
using Azure.Identity;
using Microsoft.Agents.AI;
using Microsoft.Agents.AI.DurableTask;
using Microsoft.DurableTask;
using Microsoft.DurableTask.Client;
using Microsoft.DurableTask.Client.AzureManaged;
using Microsoft.DurableTask.Worker;
using Microsoft.DurableTask.Worker.AzureManaged;
using Microsoft.Extensions.DependencyInjection;
using Microsoft.Extensions.Hosting;
using Microsoft.Extensions.Logging;
using OpenAI.Chat;
// Get the Azure OpenAI endpoint and deployment name from environment variables.
string endpoint = Environment.GetEnvironmentVariable("AZURE_OPENAI_ENDPOINT")
?? throw new InvalidOperationException("AZURE_OPENAI_ENDPOINT is not set.");
string deploymentName = Environment.GetEnvironmentVariable("AZURE_OPENAI_DEPLOYMENT")
?? throw new InvalidOperationException("AZURE_OPENAI_DEPLOYMENT is not set.");
// Get DTS connection string from environment variable
string dtsConnectionString = Environment.GetEnvironmentVariable("DURABLE_TASK_SCHEDULER_CONNECTION_STRING")
?? "Endpoint=http://localhost:8080;TaskHub=default;Authentication=None";
// Use Azure Key Credential if provided, otherwise use Azure CLI Credential.
string? azureOpenAiKey = Environment.GetEnvironmentVariable("AZURE_OPENAI_KEY");
AzureOpenAIClient client = !string.IsNullOrEmpty(azureOpenAiKey)
? new AzureOpenAIClient(new Uri(endpoint), new AzureKeyCredential(azureOpenAiKey))
: new AzureOpenAIClient(new Uri(endpoint), new AzureCliCredential());
// Two agents used by the orchestration to demonstrate concurrent execution.
const string PhysicistName = "PhysicistAgent";
const string PhysicistInstructions = "You are an expert in physics. You answer questions from a physics perspective.";
const string ChemistName = "ChemistAgent";
const string ChemistInstructions = "You are a middle school chemistry teacher. You answer questions so that middle school students can understand.";
AIAgent physicistAgent = client.GetChatClient(deploymentName).AsAIAgent(PhysicistInstructions, PhysicistName);
AIAgent chemistAgent = client.GetChatClient(deploymentName).AsAIAgent(ChemistInstructions, ChemistName);
// Orchestrator function
static async Task<object> RunOrchestratorAsync(TaskOrchestrationContext context, string prompt)
{
// Get both agents
DurableAIAgent physicist = context.GetAgent(PhysicistName);
DurableAIAgent chemist = context.GetAgent(ChemistName);
// Start both agent runs concurrently
Task<AgentResponse<TextResponse>> physicistTask = physicist.RunAsync<TextResponse>(prompt);
Task<AgentResponse<TextResponse>> chemistTask = chemist.RunAsync<TextResponse>(prompt);
// Wait for both tasks to complete using Task.WhenAll
await Task.WhenAll(physicistTask, chemistTask);
// Get the results
TextResponse physicistResponse = (await physicistTask).Result;
TextResponse chemistResponse = (await chemistTask).Result;
// Return the result as a structured, anonymous type
return new
{
physicist = physicistResponse.Text,
chemist = chemistResponse.Text,
};
}
// Configure the console app to host the AI agents.
IHost host = Host.CreateDefaultBuilder(args)
.ConfigureLogging(loggingBuilder => loggingBuilder.SetMinimumLevel(LogLevel.Warning))
.ConfigureServices(services =>
{
services.ConfigureDurableAgents(
options =>
{
options
.AddAIAgent(physicistAgent)
.AddAIAgent(chemistAgent);
},
workerBuilder: builder =>
{
builder.UseDurableTaskScheduler(dtsConnectionString);
builder.AddTasks(
registry => registry.AddOrchestratorFunc<string, object>(nameof(RunOrchestratorAsync), RunOrchestratorAsync));
},
clientBuilder: builder => builder.UseDurableTaskScheduler(dtsConnectionString));
})
.Build();
await host.StartAsync();
DurableTaskClient durableTaskClient = host.Services.GetRequiredService<DurableTaskClient>();
// Console colors for better UX
Console.ForegroundColor = ConsoleColor.Cyan;
Console.WriteLine("=== Multi-Agent Concurrent Orchestration Sample ===");
Console.ResetColor();
Console.WriteLine("Enter a question for the agents:");
Console.WriteLine();
// Read prompt from stdin
string? prompt = Console.ReadLine();
if (string.IsNullOrWhiteSpace(prompt))
{
Console.ForegroundColor = ConsoleColor.Red;
Console.Error.WriteLine("Error: Prompt is required.");
Console.ResetColor();
Environment.Exit(1);
return;
}
Console.WriteLine();
Console.ForegroundColor = ConsoleColor.Gray;
Console.WriteLine("Starting orchestration...");
Console.ResetColor();
try
{
// Start the orchestration
string instanceId = await durableTaskClient.ScheduleNewOrchestrationInstanceAsync(
orchestratorName: nameof(RunOrchestratorAsync),
input: prompt);
Console.ForegroundColor = ConsoleColor.Gray;
Console.WriteLine($"Orchestration started with instance ID: {instanceId}");
Console.WriteLine("Waiting for completion...");
Console.ResetColor();
// Wait for orchestration to complete
OrchestrationMetadata status = await durableTaskClient.WaitForInstanceCompletionAsync(
instanceId,
getInputsAndOutputs: true,
CancellationToken.None);
Console.WriteLine();
if (status.RuntimeStatus == OrchestrationRuntimeStatus.Completed)
{
Console.ForegroundColor = ConsoleColor.Green;
Console.WriteLine("✓ Orchestration completed successfully!");
Console.ResetColor();
Console.WriteLine();
// Parse the output
using JsonDocument doc = JsonDocument.Parse(status.SerializedOutput!);
JsonElement output = doc.RootElement;
Console.ForegroundColor = ConsoleColor.Yellow;
Console.WriteLine("Physicist's response:");
Console.ResetColor();
Console.WriteLine(output.GetProperty("physicist").GetString());
Console.WriteLine();
Console.ForegroundColor = ConsoleColor.Yellow;
Console.WriteLine("Chemist's response:");
Console.ResetColor();
Console.WriteLine(output.GetProperty("chemist").GetString());
}
else if (status.RuntimeStatus == OrchestrationRuntimeStatus.Failed)
{
Console.ForegroundColor = ConsoleColor.Red;
Console.WriteLine("✗ Orchestration failed!");
Console.ResetColor();
if (status.FailureDetails != null)
{
Console.WriteLine($"Error: {status.FailureDetails.ErrorMessage}");
}
Environment.Exit(1);
}
else
{
Console.ForegroundColor = ConsoleColor.Yellow;
Console.WriteLine($"Orchestration status: {status.RuntimeStatus}");
Console.ResetColor();
}
}
catch (Exception ex)
{
Console.ForegroundColor = ConsoleColor.Red;
Console.Error.WriteLine($"Error: {ex.Message}");
Console.ResetColor();
Environment.Exit(1);
}
finally
{
await host.StopAsync();
}
@@ -0,0 +1,68 @@
# Multi-Agent Concurrent Orchestration Sample
This sample demonstrates how to use the durable agents extension to create a console app that orchestrates concurrent execution of multiple AI agents using durable orchestration.
## Key Concepts Demonstrated
- Running multiple agents concurrently in a single orchestration
- Using `Task.WhenAll` to wait for concurrent agent executions
- Combining results from multiple agents into a single response
- Waiting for orchestration completion using `WaitForInstanceCompletionAsync`
## Environment Setup
See the [README.md](../README.md) file in the parent directory for more information on how to configure the environment, including how to install and run common sample dependencies.
## Running the Sample
With the environment setup, you can run the sample:
```bash
cd dotnet/samples/DurableAgents/ConsoleApps/03_AgentOrchestration_Concurrency
dotnet run --framework net10.0
```
The app will prompt you for a question:
```text
=== Multi-Agent Concurrent Orchestration Sample ===
Enter a question for the agents:
What is temperature?
```
The orchestration will run both agents concurrently and display their responses:
```text
Orchestration started with instance ID: 86313f1d45fb42eeb50b1852626bf3ff
Waiting for completion...
✓ Orchestration completed successfully!
Physicist's response:
Temperature is a measure of the average kinetic energy of particles in a system...
Chemist's response:
From a chemistry perspective, temperature is crucial for chemical reactions...
```
Both agents run in parallel, and the orchestration waits for both to complete before returning the combined results.
## Viewing Orchestration State
You can view the state of the orchestration in the Durable Task Scheduler dashboard:
1. Open your browser and navigate to `http://localhost:8082`
2. In the dashboard, you can see:
- **Orchestrations**: View the orchestration instance, including its runtime status, input, output, and execution history
- **Agents**: View the state of both the PhysicistAgent and ChemistAgent, including their individual conversation histories
The orchestration instance ID is displayed in the console output. You can use this ID to find the specific orchestration in the dashboard and inspect how the concurrent agent executions were coordinated, including the timing of when each agent started and completed.
## Scriptable Usage
You can also pipe input to the app:
```bash
echo "What is temperature?" | dotnet run
```
@@ -0,0 +1,30 @@
<Project Sdk="Microsoft.NET.Sdk">
<PropertyGroup>
<TargetFrameworks>net10.0</TargetFrameworks>
<OutputType>Exe</OutputType>
<ImplicitUsings>enable</ImplicitUsings>
<Nullable>enable</Nullable>
<AssemblyName>AgentOrchestration_Conditionals</AssemblyName>
<RootNamespace>AgentOrchestration_Conditionals</RootNamespace>
</PropertyGroup>
<ItemGroup>
<PackageReference Include="Azure.AI.OpenAI" />
<PackageReference Include="Azure.Identity" />
<PackageReference Include="Microsoft.DurableTask.Client.AzureManaged" />
<PackageReference Include="Microsoft.DurableTask.Worker.AzureManaged" />
<PackageReference Include="Microsoft.Extensions.Hosting" />
</ItemGroup>
<!-- Local projects that should be switched to package references when using the sample outside of this MAF repo -->
<!--
<ItemGroup>
<PackageReference Include="Microsoft.Agents.AI.DurableTask" />
<PackageReference Include="Microsoft.Agents.AI.OpenAI" />
</ItemGroup>
-->
<ItemGroup>
<ProjectReference Include="..\..\..\..\src\Microsoft.Agents.AI.DurableTask\Microsoft.Agents.AI.DurableTask.csproj" />
<ProjectReference Include="..\..\..\..\src\Microsoft.Agents.AI.OpenAI\Microsoft.Agents.AI.OpenAI.csproj" />
</ItemGroup>
</Project>
@@ -0,0 +1,38 @@
// Copyright (c) Microsoft. All rights reserved.
using System.Text.Json.Serialization;
namespace AgentOrchestration_Conditionals;
/// <summary>
/// Represents an email input for spam detection and response generation.
/// </summary>
public sealed class Email
{
[JsonPropertyName("email_id")]
public string EmailId { get; set; } = string.Empty;
[JsonPropertyName("email_content")]
public string EmailContent { get; set; } = string.Empty;
}
/// <summary>
/// Represents the result of spam detection analysis.
/// </summary>
public sealed class DetectionResult
{
[JsonPropertyName("is_spam")]
public bool IsSpam { get; set; }
[JsonPropertyName("reason")]
public string Reason { get; set; } = string.Empty;
}
/// <summary>
/// Represents a generated email response.
/// </summary>
public sealed class EmailResponse
{
[JsonPropertyName("response")]
public string Response { get; set; } = string.Empty;
}
@@ -0,0 +1,228 @@
// Copyright (c) Microsoft. All rights reserved.
using AgentOrchestration_Conditionals;
using Azure;
using Azure.AI.OpenAI;
using Azure.Identity;
using Microsoft.Agents.AI;
using Microsoft.Agents.AI.DurableTask;
using Microsoft.DurableTask;
using Microsoft.DurableTask.Client;
using Microsoft.DurableTask.Client.AzureManaged;
using Microsoft.DurableTask.Worker;
using Microsoft.DurableTask.Worker.AzureManaged;
using Microsoft.Extensions.DependencyInjection;
using Microsoft.Extensions.Hosting;
using Microsoft.Extensions.Logging;
using OpenAI.Chat;
// Get the Azure OpenAI endpoint and deployment name from environment variables.
string endpoint = Environment.GetEnvironmentVariable("AZURE_OPENAI_ENDPOINT")
?? throw new InvalidOperationException("AZURE_OPENAI_ENDPOINT is not set.");
string deploymentName = Environment.GetEnvironmentVariable("AZURE_OPENAI_DEPLOYMENT")
?? throw new InvalidOperationException("AZURE_OPENAI_DEPLOYMENT is not set.");
// Get DTS connection string from environment variable
string dtsConnectionString = Environment.GetEnvironmentVariable("DURABLE_TASK_SCHEDULER_CONNECTION_STRING")
?? "Endpoint=http://localhost:8080;TaskHub=default;Authentication=None";
// Use Azure Key Credential if provided, otherwise use Azure CLI Credential.
string? azureOpenAiKey = Environment.GetEnvironmentVariable("AZURE_OPENAI_KEY");
AzureOpenAIClient client = !string.IsNullOrEmpty(azureOpenAiKey)
? new AzureOpenAIClient(new Uri(endpoint), new AzureKeyCredential(azureOpenAiKey))
: new AzureOpenAIClient(new Uri(endpoint), new AzureCliCredential());
// Spam detection agent
const string SpamDetectionAgentName = "SpamDetectionAgent";
const string SpamDetectionAgentInstructions =
"""
You are an expert email spam detection system. Analyze emails and determine if they are spam.
Return your analysis as JSON with 'is_spam' (boolean) and 'reason' (string) fields.
""";
// Email assistant agent
const string EmailAssistantAgentName = "EmailAssistantAgent";
const string EmailAssistantAgentInstructions =
"""
You are a professional email assistant. Draft professional, courteous, and helpful email responses.
Return your response as JSON with a 'response' field containing the reply.
""";
AIAgent spamDetectionAgent = client.GetChatClient(deploymentName).AsAIAgent(SpamDetectionAgentInstructions, SpamDetectionAgentName);
AIAgent emailAssistantAgent = client.GetChatClient(deploymentName).AsAIAgent(EmailAssistantAgentInstructions, EmailAssistantAgentName);
// Orchestrator function
static async Task<string> RunOrchestratorAsync(TaskOrchestrationContext context, Email email)
{
// Get the spam detection agent
DurableAIAgent spamDetectionAgent = context.GetAgent(SpamDetectionAgentName);
AgentThread spamThread = await spamDetectionAgent.GetNewThreadAsync();
// Step 1: Check if the email is spam
AgentResponse<DetectionResult> spamDetectionResponse = await spamDetectionAgent.RunAsync<DetectionResult>(
message:
$"""
Analyze this email for spam content and return a JSON response with 'is_spam' (boolean) and 'reason' (string) fields:
Email ID: {email.EmailId}
Content: {email.EmailContent}
""",
thread: spamThread);
DetectionResult result = spamDetectionResponse.Result;
// Step 2: Conditional logic based on spam detection result
if (result.IsSpam)
{
// Handle spam email
return await context.CallActivityAsync<string>(nameof(HandleSpamEmail), result.Reason);
}
// Generate and send response for legitimate email
DurableAIAgent emailAssistantAgent = context.GetAgent(EmailAssistantAgentName);
AgentThread emailThread = await emailAssistantAgent.GetNewThreadAsync();
AgentResponse<EmailResponse> emailAssistantResponse = await emailAssistantAgent.RunAsync<EmailResponse>(
message:
$"""
Draft a professional response to this email. Return a JSON response with a 'response' field containing the reply:
Email ID: {email.EmailId}
Content: {email.EmailContent}
""",
thread: emailThread);
EmailResponse emailResponse = emailAssistantResponse.Result;
return await context.CallActivityAsync<string>(nameof(SendEmail), emailResponse.Response);
}
// Activity functions
static void HandleSpamEmail(TaskActivityContext context, string reason)
{
Console.WriteLine($"Email marked as spam: {reason}");
}
static void SendEmail(TaskActivityContext context, string message)
{
Console.WriteLine($"Email sent: {message}");
}
// Configure the console app to host the AI agents.
IHost host = Host.CreateDefaultBuilder(args)
.ConfigureLogging(loggingBuilder => loggingBuilder.SetMinimumLevel(LogLevel.Warning))
.ConfigureServices(services =>
{
services.ConfigureDurableAgents(
options =>
{
options
.AddAIAgent(spamDetectionAgent)
.AddAIAgent(emailAssistantAgent);
},
workerBuilder: builder =>
{
builder.UseDurableTaskScheduler(dtsConnectionString);
builder.AddTasks(registry =>
{
registry.AddOrchestratorFunc<Email>(nameof(RunOrchestratorAsync), RunOrchestratorAsync);
registry.AddActivityFunc<string>(nameof(HandleSpamEmail), HandleSpamEmail);
registry.AddActivityFunc<string>(nameof(SendEmail), SendEmail);
});
},
clientBuilder: builder => builder.UseDurableTaskScheduler(dtsConnectionString));
})
.Build();
await host.StartAsync();
DurableTaskClient durableTaskClient = host.Services.GetRequiredService<DurableTaskClient>();
// Console colors for better UX
Console.ForegroundColor = ConsoleColor.Cyan;
Console.WriteLine("=== Multi-Agent Conditional Orchestration Sample ===");
Console.ResetColor();
Console.WriteLine("Enter email content:");
Console.WriteLine();
// Read email content from stdin
string? emailContent = Console.ReadLine();
if (string.IsNullOrWhiteSpace(emailContent))
{
Console.ForegroundColor = ConsoleColor.Red;
Console.Error.WriteLine("Error: Email content is required.");
Console.ResetColor();
Environment.Exit(1);
return;
}
// Generate email ID automatically
Email email = new()
{
EmailId = $"email-{Guid.NewGuid():N}",
EmailContent = emailContent
};
Console.WriteLine();
Console.ForegroundColor = ConsoleColor.Gray;
Console.WriteLine("Starting orchestration...");
Console.ResetColor();
try
{
// Start the orchestration
string instanceId = await durableTaskClient.ScheduleNewOrchestrationInstanceAsync(
orchestratorName: nameof(RunOrchestratorAsync),
input: email);
Console.ForegroundColor = ConsoleColor.Gray;
Console.WriteLine($"Orchestration started with instance ID: {instanceId}");
Console.WriteLine("Waiting for completion...");
Console.ResetColor();
// Wait for orchestration to complete
OrchestrationMetadata status = await durableTaskClient.WaitForInstanceCompletionAsync(
instanceId,
getInputsAndOutputs: true,
CancellationToken.None);
Console.WriteLine();
if (status.RuntimeStatus == OrchestrationRuntimeStatus.Completed)
{
Console.ForegroundColor = ConsoleColor.Green;
Console.WriteLine("✓ Orchestration completed successfully!");
Console.ResetColor();
Console.WriteLine();
Console.ForegroundColor = ConsoleColor.Yellow;
Console.Write("Result: ");
Console.ResetColor();
Console.WriteLine(status.ReadOutputAs<string>());
}
else if (status.RuntimeStatus == OrchestrationRuntimeStatus.Failed)
{
Console.ForegroundColor = ConsoleColor.Red;
Console.WriteLine("✗ Orchestration failed!");
Console.ResetColor();
if (status.FailureDetails != null)
{
Console.WriteLine($"Error: {status.FailureDetails.ErrorMessage}");
}
Environment.Exit(1);
}
else
{
Console.ForegroundColor = ConsoleColor.Yellow;
Console.WriteLine($"Orchestration status: {status.RuntimeStatus}");
Console.ResetColor();
}
}
catch (Exception ex)
{
Console.ForegroundColor = ConsoleColor.Red;
Console.Error.WriteLine($"Error: {ex.Message}");
Console.ResetColor();
Environment.Exit(1);
}
finally
{
await host.StopAsync();
}
@@ -0,0 +1,95 @@
# Multi-Agent Conditional Orchestration Sample
This sample demonstrates how to use the durable agents extension to create a console app that orchestrates multiple AI agents with conditional logic based on the results of previous agent interactions.
## Key Concepts Demonstrated
- Multi-agent orchestration with conditional branching
- Using agent responses to determine workflow paths
- Activity functions for non-agent operations
- Waiting for orchestration completion using `WaitForInstanceCompletionAsync`
## Environment Setup
See the [README.md](../README.md) file in the parent directory for more information on how to configure the environment, including how to install and run common sample dependencies.
## Running the Sample
With the environment setup, you can run the sample:
```bash
cd dotnet/samples/DurableAgents/ConsoleApps/04_AgentOrchestration_Conditionals
dotnet run --framework net10.0
```
The app will prompt you for email content. You can test both legitimate emails and spam emails:
### Testing with a Legitimate Email
```text
=== Multi-Agent Conditional Orchestration Sample ===
Enter email content:
Hi John, I hope you're doing well. I wanted to follow up on our meeting yesterday about the quarterly report. Could you please send me the updated figures by Friday? Thanks!
```
The orchestration will analyze the email and display the result:
```text
Orchestration started with instance ID: 86313f1d45fb42eeb50b1852626bf3ff
Waiting for completion...
✓ Orchestration completed successfully!
Result: Email sent: Thank you for your email. I'll prepare the updated figures...
```
### Testing with a Spam Email
```text
=== Multi-Agent Conditional Orchestration Sample ===
Enter email content:
URGENT! You've won $1,000,000! Click here now to claim your prize! Limited time offer! Don't miss out!
```
The orchestration will detect it as spam and display:
```text
Orchestration started with instance ID: 86313f1d45fb42eeb50b1852626bf3ff
Waiting for completion...
✓ Orchestration completed successfully!
Result: Email marked as spam: Contains suspicious claims about winning money and urgent action requests...
```
## Scriptable Usage
You can also pipe email content to the app:
```bash
# Test with a legitimate email
echo "Hi John, I hope you're doing well..." | dotnet run
# Test with a spam email
echo "URGENT! You've won $1,000,000! Click here now!" | dotnet run
```
The orchestration will proceed as follows:
1. The SpamDetectionAgent analyzes the email to determine if it's spam
2. Based on the result:
- If spam: The orchestration calls the `HandleSpamEmail` activity function
- If not spam: The EmailAssistantAgent drafts a response, then the `SendEmail` activity function is called
## Viewing Orchestration State
You can view the state of the orchestration in the Durable Task Scheduler dashboard:
1. Open your browser and navigate to `http://localhost:8082`
2. In the dashboard, you can see:
- **Orchestrations**: View the orchestration instance, including its runtime status, input, output, and execution history
- **Agents**: View the state of both the SpamDetectionAgent and EmailAssistantAgent
The orchestration instance ID is displayed in the console output. You can use this ID to find the specific orchestration in the dashboard and inspect the conditional branching logic, including which path was taken based on the spam detection result.
@@ -0,0 +1,30 @@
<Project Sdk="Microsoft.NET.Sdk">
<PropertyGroup>
<TargetFrameworks>net10.0</TargetFrameworks>
<OutputType>Exe</OutputType>
<ImplicitUsings>enable</ImplicitUsings>
<Nullable>enable</Nullable>
<AssemblyName>AgentOrchestration_HITL</AssemblyName>
<RootNamespace>AgentOrchestration_HITL</RootNamespace>
</PropertyGroup>
<ItemGroup>
<PackageReference Include="Azure.AI.OpenAI" />
<PackageReference Include="Azure.Identity" />
<PackageReference Include="Microsoft.DurableTask.Client.AzureManaged" />
<PackageReference Include="Microsoft.DurableTask.Worker.AzureManaged" />
<PackageReference Include="Microsoft.Extensions.Hosting" />
</ItemGroup>
<!-- Local projects that should be switched to package references when using the sample outside of this MAF repo -->
<!--
<ItemGroup>
<PackageReference Include="Microsoft.Agents.AI.DurableTask" />
<PackageReference Include="Microsoft.Agents.AI.OpenAI" />
</ItemGroup>
-->
<ItemGroup>
<ProjectReference Include="..\..\..\..\src\Microsoft.Agents.AI.DurableTask\Microsoft.Agents.AI.DurableTask.csproj" />
<ProjectReference Include="..\..\..\..\src\Microsoft.Agents.AI.OpenAI\Microsoft.Agents.AI.OpenAI.csproj" />
</ItemGroup>
</Project>
@@ -0,0 +1,44 @@
// Copyright (c) Microsoft. All rights reserved.
using System.Text.Json.Serialization;
namespace AgentOrchestration_HITL;
/// <summary>
/// Represents the input for the Human-in-the-Loop content generation workflow.
/// </summary>
public sealed class ContentGenerationInput
{
[JsonPropertyName("topic")]
public string Topic { get; set; } = string.Empty;
[JsonPropertyName("max_review_attempts")]
public int MaxReviewAttempts { get; set; } = 3;
[JsonPropertyName("approval_timeout_hours")]
public float ApprovalTimeoutHours { get; set; } = 72;
}
/// <summary>
/// Represents the content generated by the writer agent.
/// </summary>
public sealed class GeneratedContent
{
[JsonPropertyName("title")]
public string Title { get; set; } = string.Empty;
[JsonPropertyName("content")]
public string Content { get; set; } = string.Empty;
}
/// <summary>
/// Represents the human approval response.
/// </summary>
public sealed class HumanApprovalResponse
{
[JsonPropertyName("approved")]
public bool Approved { get; set; }
[JsonPropertyName("feedback")]
public string Feedback { get; set; } = string.Empty;
}
@@ -0,0 +1,333 @@
// Copyright (c) Microsoft. All rights reserved.
using System.Text.Json;
using AgentOrchestration_HITL;
using Azure;
using Azure.AI.OpenAI;
using Azure.Identity;
using Microsoft.Agents.AI;
using Microsoft.Agents.AI.DurableTask;
using Microsoft.DurableTask;
using Microsoft.DurableTask.Client;
using Microsoft.DurableTask.Client.AzureManaged;
using Microsoft.DurableTask.Worker;
using Microsoft.DurableTask.Worker.AzureManaged;
using Microsoft.Extensions.DependencyInjection;
using Microsoft.Extensions.Hosting;
using Microsoft.Extensions.Logging;
using OpenAI.Chat;
// Get the Azure OpenAI endpoint and deployment name from environment variables.
string endpoint = Environment.GetEnvironmentVariable("AZURE_OPENAI_ENDPOINT")
?? throw new InvalidOperationException("AZURE_OPENAI_ENDPOINT is not set.");
string deploymentName = Environment.GetEnvironmentVariable("AZURE_OPENAI_DEPLOYMENT")
?? throw new InvalidOperationException("AZURE_OPENAI_DEPLOYMENT is not set.");
// Get DTS connection string from environment variable
string dtsConnectionString = Environment.GetEnvironmentVariable("DURABLE_TASK_SCHEDULER_CONNECTION_STRING")
?? "Endpoint=http://localhost:8080;TaskHub=default;Authentication=None";
// Use Azure Key Credential if provided, otherwise use Azure CLI Credential.
string? azureOpenAiKey = Environment.GetEnvironmentVariable("AZURE_OPENAI_KEY");
AzureOpenAIClient client = !string.IsNullOrEmpty(azureOpenAiKey)
? new AzureOpenAIClient(new Uri(endpoint), new AzureKeyCredential(azureOpenAiKey))
: new AzureOpenAIClient(new Uri(endpoint), new AzureCliCredential());
// Single agent used by the orchestration to demonstrate human-in-the-loop workflow.
const string WriterName = "WriterAgent";
const string WriterInstructions =
"""
You are a professional content writer who creates high-quality articles on various topics.
You write engaging, informative, and well-structured content that follows best practices for readability and accuracy.
""";
AIAgent writerAgent = client.GetChatClient(deploymentName).AsAIAgent(WriterInstructions, WriterName);
// Orchestrator function
static async Task<object> RunOrchestratorAsync(TaskOrchestrationContext context, ContentGenerationInput input)
{
// Get the writer agent
DurableAIAgent writerAgent = context.GetAgent("WriterAgent");
AgentThread writerThread = await writerAgent.GetNewThreadAsync();
// Set initial status
context.SetCustomStatus($"Starting content generation for topic: {input.Topic}");
// Step 1: Generate initial content
AgentResponse<GeneratedContent> writerResponse = await writerAgent.RunAsync<GeneratedContent>(
message: $"Write a short article about '{input.Topic}' in less than 300 words.",
thread: writerThread);
GeneratedContent content = writerResponse.Result;
// Human-in-the-loop iteration - we set a maximum number of attempts to avoid infinite loops
int iterationCount = 0;
while (iterationCount++ < input.MaxReviewAttempts)
{
context.SetCustomStatus(
$"Requesting human feedback. Iteration #{iterationCount}. Timeout: {input.ApprovalTimeoutHours} hour(s).");
// Step 2: Notify user to review the content
await context.CallActivityAsync(nameof(NotifyUserForApproval), content);
// Step 3: Wait for human feedback with configurable timeout
HumanApprovalResponse humanResponse;
try
{
humanResponse = await context.WaitForExternalEvent<HumanApprovalResponse>(
eventName: "HumanApproval",
timeout: TimeSpan.FromHours(input.ApprovalTimeoutHours));
}
catch (OperationCanceledException)
{
// Timeout occurred - treat as rejection
context.SetCustomStatus(
$"Human approval timed out after {input.ApprovalTimeoutHours} hour(s). Treating as rejection.");
throw new TimeoutException($"Human approval timed out after {input.ApprovalTimeoutHours} hour(s).");
}
if (humanResponse.Approved)
{
context.SetCustomStatus("Content approved by human reviewer. Publishing content...");
// Step 4: Publish the approved content
await context.CallActivityAsync(nameof(PublishContent), content);
context.SetCustomStatus($"Content published successfully at {context.CurrentUtcDateTime:s}");
return new { content = content.Content };
}
context.SetCustomStatus("Content rejected by human reviewer. Incorporating feedback and regenerating...");
// Incorporate human feedback and regenerate
writerResponse = await writerAgent.RunAsync<GeneratedContent>(
message: $"""
The content was rejected by a human reviewer. Please rewrite the article incorporating their feedback.
Human Feedback: {humanResponse.Feedback}
""",
thread: writerThread);
content = writerResponse.Result;
}
// If we reach here, it means we exhausted the maximum number of iterations
throw new InvalidOperationException(
$"Content could not be approved after {input.MaxReviewAttempts} iterations.");
}
// Activity functions
static void NotifyUserForApproval(TaskActivityContext context, GeneratedContent content)
{
// In a real implementation, this would send notifications via email, SMS, etc.
Console.WriteLine(
$"""
NOTIFICATION: Please review the following content for approval:
Title: {content.Title}
Content: {content.Content}
Use the approval endpoint to approve or reject this content.
""");
}
static void PublishContent(TaskActivityContext context, GeneratedContent content)
{
// In a real implementation, this would publish to a CMS, website, etc.
Console.WriteLine(
$"""
PUBLISHING: Content has been published successfully.
Title: {content.Title}
Content: {content.Content}
""");
}
// Configure the console app to host the AI agent.
IHost host = Host.CreateDefaultBuilder(args)
.ConfigureLogging(loggingBuilder => loggingBuilder.SetMinimumLevel(LogLevel.Warning))
.ConfigureServices(services =>
{
services.ConfigureDurableAgents(
options => options.AddAIAgent(writerAgent),
workerBuilder: builder =>
{
builder.UseDurableTaskScheduler(dtsConnectionString);
builder.AddTasks(registry =>
{
registry.AddOrchestratorFunc<ContentGenerationInput>(nameof(RunOrchestratorAsync), RunOrchestratorAsync);
registry.AddActivityFunc<GeneratedContent>(nameof(NotifyUserForApproval), NotifyUserForApproval);
registry.AddActivityFunc<GeneratedContent>(nameof(PublishContent), PublishContent);
});
},
clientBuilder: builder => builder.UseDurableTaskScheduler(dtsConnectionString));
})
.Build();
await host.StartAsync();
DurableTaskClient durableTaskClient = host.Services.GetRequiredService<DurableTaskClient>();
// Console colors for better UX
Console.ForegroundColor = ConsoleColor.Cyan;
Console.WriteLine("=== Human-in-the-Loop Orchestration Sample ===");
Console.ResetColor();
Console.WriteLine("Enter topic for content generation:");
Console.WriteLine();
// Read topic from stdin
string? topic = Console.ReadLine();
if (string.IsNullOrWhiteSpace(topic))
{
Console.ForegroundColor = ConsoleColor.Red;
Console.Error.WriteLine("Error: Topic is required.");
Console.ResetColor();
Environment.Exit(1);
return;
}
// Prompt for optional parameters with defaults
Console.WriteLine();
Console.WriteLine("Max review attempts (default: 3):");
string? maxAttemptsInput = Console.ReadLine();
int maxReviewAttempts = int.TryParse(maxAttemptsInput, out int maxAttempts) && maxAttempts > 0
? maxAttempts
: 3;
Console.WriteLine("Approval timeout in hours (default: 72):");
string? timeoutInput = Console.ReadLine();
float approvalTimeoutHours = float.TryParse(timeoutInput, out float timeout) && timeout > 0
? timeout
: 72;
ContentGenerationInput input = new()
{
Topic = topic,
MaxReviewAttempts = maxReviewAttempts,
ApprovalTimeoutHours = approvalTimeoutHours
};
Console.WriteLine();
Console.ForegroundColor = ConsoleColor.Gray;
Console.WriteLine("Starting orchestration...");
Console.ResetColor();
try
{
// Start the orchestration
string instanceId = await durableTaskClient.ScheduleNewOrchestrationInstanceAsync(
orchestratorName: nameof(RunOrchestratorAsync),
input: input);
Console.ForegroundColor = ConsoleColor.Gray;
Console.WriteLine($"Orchestration started with instance ID: {instanceId}");
Console.WriteLine("Waiting for human approval...");
Console.ResetColor();
Console.WriteLine();
// Monitor orchestration status and handle approval prompts
using CancellationTokenSource cts = new();
Task orchestrationTask = Task.Run(async () =>
{
while (!cts.Token.IsCancellationRequested)
{
OrchestrationMetadata? status = await durableTaskClient.GetInstanceAsync(
instanceId,
getInputsAndOutputs: true,
cts.Token);
if (status == null)
{
await Task.Delay(TimeSpan.FromSeconds(1), cts.Token);
continue;
}
// Check if we're waiting for approval
if (status.SerializedCustomStatus != null)
{
string? customStatus = status.ReadCustomStatusAs<string>();
if (customStatus?.StartsWith("Requesting human feedback", StringComparison.OrdinalIgnoreCase) == true)
{
// Prompt user for approval
Console.ForegroundColor = ConsoleColor.Yellow;
Console.WriteLine("Content is ready for review. Check the logs above for details.");
Console.Write("Approve? (y/n): ");
Console.ResetColor();
string? approvalInput = Console.ReadLine();
bool approved = approvalInput?.Trim().Equals("y", StringComparison.OrdinalIgnoreCase) == true;
Console.Write("Feedback (optional): ");
string? feedback = Console.ReadLine() ?? "";
HumanApprovalResponse approvalResponse = new()
{
Approved = approved,
Feedback = feedback
};
await durableTaskClient.RaiseEventAsync(instanceId, "HumanApproval", approvalResponse);
}
}
if (status.RuntimeStatus is OrchestrationRuntimeStatus.Completed or OrchestrationRuntimeStatus.Failed or OrchestrationRuntimeStatus.Terminated)
{
break;
}
await Task.Delay(TimeSpan.FromSeconds(1), cts.Token);
}
}, cts.Token);
// Wait for orchestration to complete
OrchestrationMetadata finalStatus = await durableTaskClient.WaitForInstanceCompletionAsync(
instanceId,
getInputsAndOutputs: true,
CancellationToken.None);
cts.Cancel();
await orchestrationTask;
Console.WriteLine();
if (finalStatus.RuntimeStatus == OrchestrationRuntimeStatus.Completed)
{
Console.ForegroundColor = ConsoleColor.Green;
Console.WriteLine("✓ Orchestration completed successfully!");
Console.ResetColor();
Console.WriteLine();
JsonElement output = finalStatus.ReadOutputAs<JsonElement>();
if (output.TryGetProperty("content", out JsonElement contentElement))
{
Console.ForegroundColor = ConsoleColor.Yellow;
Console.WriteLine("Published content:");
Console.ResetColor();
Console.WriteLine(contentElement.GetString());
}
}
else if (finalStatus.RuntimeStatus == OrchestrationRuntimeStatus.Failed)
{
Console.ForegroundColor = ConsoleColor.Red;
Console.WriteLine("✗ Orchestration failed!");
Console.ResetColor();
if (finalStatus.FailureDetails != null)
{
Console.WriteLine($"Error: {finalStatus.FailureDetails.ErrorMessage}");
}
Environment.Exit(1);
}
else
{
Console.ForegroundColor = ConsoleColor.Yellow;
Console.WriteLine($"Orchestration status: {finalStatus.RuntimeStatus}");
Console.ResetColor();
}
}
catch (Exception ex)
{
Console.ForegroundColor = ConsoleColor.Red;
Console.Error.WriteLine($"Error: {ex.Message}");
Console.ResetColor();
Environment.Exit(1);
}
finally
{
await host.StopAsync();
}
@@ -0,0 +1,73 @@
# Human-in-the-Loop Orchestration Sample
This sample demonstrates how to use the durable agents extension to create a console app that implements a human-in-the-loop workflow using durable orchestration, including interactive approval prompts.
## Key Concepts Demonstrated
- Human-in-the-loop workflows with durable orchestration
- External event handling for human approval/rejection
- Timeout handling for approval requests
- Iterative content refinement based on human feedback
## Environment Setup
See the [README.md](../README.md) file in the parent directory for more information on how to configure the environment, including how to install and run common sample dependencies.
## Running the Sample
With the environment setup, you can run the sample:
```bash
cd dotnet/samples/DurableAgents/ConsoleApps/05_AgentOrchestration_HITL
dotnet run --framework net10.0
```
The app will prompt you for input:
```text
=== Human-in-the-Loop Orchestration Sample ===
Enter topic for content generation:
The Future of Artificial Intelligence
Max review attempts (default: 3):
3
Approval timeout in hours (default: 72):
72
```
The orchestration will generate content and prompt you for approval:
```text
Orchestration started with instance ID: 86313f1d45fb42eeb50b1852626bf3ff
=== NOTIFICATION: Content Ready for Review ===
Title: The Future of Artificial Intelligence
Content:
[Generated content appears here]
Please review the content above and provide your approval.
Content is ready for review. Check the logs above for details.
Approve? (y/n): n
Feedback (optional): Please add more details about the ethical implications.
```
The orchestration will incorporate your feedback and regenerate the content. Once approved, it will publish and complete.
## Viewing Orchestration State
You can view the state of the orchestration in the Durable Task Scheduler dashboard:
1. Open your browser and navigate to `http://localhost:8082`
2. In the dashboard, you can see:
- **Orchestrations**: View the orchestration instance, including its runtime status, custom status (which shows approval state), input, output, and execution history
- **Agents**: View the state of the WriterAgent, including conversation history
The orchestration instance ID is displayed in the console output. You can use this ID to find the specific orchestration in the dashboard and inspect:
- The custom status field, which shows the current state of the approval workflow
- When the orchestration is waiting for external events
- The iteration count and feedback history
- The final published content
@@ -0,0 +1,30 @@
<Project Sdk="Microsoft.NET.Sdk">
<PropertyGroup>
<TargetFrameworks>net10.0</TargetFrameworks>
<OutputType>Exe</OutputType>
<ImplicitUsings>enable</ImplicitUsings>
<Nullable>enable</Nullable>
<AssemblyName>LongRunningTools</AssemblyName>
<RootNamespace>LongRunningTools</RootNamespace>
</PropertyGroup>
<ItemGroup>
<PackageReference Include="Azure.AI.OpenAI" />
<PackageReference Include="Azure.Identity" />
<PackageReference Include="Microsoft.DurableTask.Client.AzureManaged" />
<PackageReference Include="Microsoft.DurableTask.Worker.AzureManaged" />
<PackageReference Include="Microsoft.Extensions.Hosting" />
</ItemGroup>
<!-- Local projects that should be switched to package references when using the sample outside of this MAF repo -->
<!--
<ItemGroup>
<PackageReference Include="Microsoft.Agents.AI.DurableTask" />
<PackageReference Include="Microsoft.Agents.AI.OpenAI" />
</ItemGroup>
-->
<ItemGroup>
<ProjectReference Include="..\..\..\..\src\Microsoft.Agents.AI.DurableTask\Microsoft.Agents.AI.DurableTask.csproj" />
<ProjectReference Include="..\..\..\..\src\Microsoft.Agents.AI.OpenAI\Microsoft.Agents.AI.OpenAI.csproj" />
</ItemGroup>
</Project>
@@ -0,0 +1,44 @@
// Copyright (c) Microsoft. All rights reserved.
using System.Text.Json.Serialization;
namespace LongRunningTools;
/// <summary>
/// Represents the input for the content generation workflow.
/// </summary>
public sealed class ContentGenerationInput
{
[JsonPropertyName("topic")]
public string Topic { get; set; } = string.Empty;
[JsonPropertyName("max_review_attempts")]
public int MaxReviewAttempts { get; set; } = 3;
[JsonPropertyName("approval_timeout_hours")]
public float ApprovalTimeoutHours { get; set; } = 72;
}
/// <summary>
/// Represents the content generated by the writer agent.
/// </summary>
public sealed class GeneratedContent
{
[JsonPropertyName("title")]
public string Title { get; set; } = string.Empty;
[JsonPropertyName("content")]
public string Content { get; set; } = string.Empty;
}
/// <summary>
/// Represents the human feedback response.
/// </summary>
public sealed class HumanFeedbackResponse
{
[JsonPropertyName("approved")]
public bool Approved { get; set; }
[JsonPropertyName("feedback")]
public string Feedback { get; set; } = string.Empty;
}
@@ -0,0 +1,351 @@
// Copyright (c) Microsoft. All rights reserved.
using System.ComponentModel;
using Azure;
using Azure.AI.OpenAI;
using Azure.Identity;
using LongRunningTools;
using Microsoft.Agents.AI;
using Microsoft.Agents.AI.DurableTask;
using Microsoft.DurableTask;
using Microsoft.DurableTask.Client;
using Microsoft.DurableTask.Client.AzureManaged;
using Microsoft.DurableTask.Worker;
using Microsoft.DurableTask.Worker.AzureManaged;
using Microsoft.Extensions.AI;
using Microsoft.Extensions.DependencyInjection;
using Microsoft.Extensions.Hosting;
using Microsoft.Extensions.Logging;
using OpenAI.Chat;
// Get the Azure OpenAI endpoint and deployment name from environment variables.
string endpoint = Environment.GetEnvironmentVariable("AZURE_OPENAI_ENDPOINT")
?? throw new InvalidOperationException("AZURE_OPENAI_ENDPOINT is not set.");
string deploymentName = Environment.GetEnvironmentVariable("AZURE_OPENAI_DEPLOYMENT")
?? throw new InvalidOperationException("AZURE_OPENAI_DEPLOYMENT is not set.");
// Get DTS connection string from environment variable
string dtsConnectionString = Environment.GetEnvironmentVariable("DURABLE_TASK_SCHEDULER_CONNECTION_STRING")
?? "Endpoint=http://localhost:8080;TaskHub=default;Authentication=None";
// Use Azure Key Credential if provided, otherwise use Azure CLI Credential.
string? azureOpenAiKey = Environment.GetEnvironmentVariable("AZURE_OPENAI_KEY");
AzureOpenAIClient client = !string.IsNullOrEmpty(azureOpenAiKey)
? new AzureOpenAIClient(new Uri(endpoint), new AzureKeyCredential(azureOpenAiKey))
: new AzureOpenAIClient(new Uri(endpoint), new AzureCliCredential());
// Agent used by the orchestration to write content.
const string WriterAgentName = "Writer";
const string WriterAgentInstructions =
"""
You are a professional content writer who creates high-quality articles on various topics.
You write engaging, informative, and well-structured content that follows best practices for readability and accuracy.
""";
AIAgent writerAgent = client.GetChatClient(deploymentName).AsAIAgent(WriterAgentInstructions, WriterAgentName);
// Agent that can start content generation workflows using tools
const string PublisherAgentName = "Publisher";
const string PublisherAgentInstructions =
"""
You are a publishing agent that can manage content generation workflows.
You have access to tools to start, monitor, and raise events for content generation workflows.
""";
const string HumanFeedbackEventName = "HumanFeedback";
// Orchestrator function
static async Task<object> RunOrchestratorAsync(TaskOrchestrationContext context, ContentGenerationInput input)
{
// Get the writer agent
DurableAIAgent writerAgent = context.GetAgent(WriterAgentName);
AgentThread writerThread = await writerAgent.GetNewThreadAsync();
// Set initial status
context.SetCustomStatus($"Starting content generation for topic: {input.Topic}");
// Step 1: Generate initial content
AgentResponse<GeneratedContent> writerResponse = await writerAgent.RunAsync<GeneratedContent>(
message: $"Write a short article about '{input.Topic}'.",
thread: writerThread);
GeneratedContent content = writerResponse.Result;
// Human-in-the-loop iteration - we set a maximum number of attempts to avoid infinite loops
int iterationCount = 0;
while (iterationCount++ < input.MaxReviewAttempts)
{
context.SetCustomStatus(
new
{
message = "Requesting human feedback.",
approvalTimeoutHours = input.ApprovalTimeoutHours,
iterationCount,
content
});
// Step 2: Notify user to review the content
await context.CallActivityAsync(nameof(NotifyUserForApproval), content);
// Step 3: Wait for human feedback with configurable timeout
HumanFeedbackResponse humanResponse;
try
{
humanResponse = await context.WaitForExternalEvent<HumanFeedbackResponse>(
eventName: HumanFeedbackEventName,
timeout: TimeSpan.FromHours(input.ApprovalTimeoutHours));
}
catch (OperationCanceledException)
{
// Timeout occurred - treat as rejection
context.SetCustomStatus(
new
{
message = $"Human approval timed out after {input.ApprovalTimeoutHours} hour(s). Treating as rejection.",
iterationCount,
content
});
throw new TimeoutException($"Human approval timed out after {input.ApprovalTimeoutHours} hour(s).");
}
if (humanResponse.Approved)
{
context.SetCustomStatus(new
{
message = "Content approved by human reviewer. Publishing content...",
content
});
// Step 4: Publish the approved content
await context.CallActivityAsync(nameof(PublishContent), content);
context.SetCustomStatus(new
{
message = $"Content published successfully at {context.CurrentUtcDateTime:s}",
humanFeedback = humanResponse,
content
});
return new { content = content.Content };
}
context.SetCustomStatus(new
{
message = "Content rejected by human reviewer. Incorporating feedback and regenerating...",
humanFeedback = humanResponse,
content
});
// Incorporate human feedback and regenerate
writerResponse = await writerAgent.RunAsync<GeneratedContent>(
message: $"""
The content was rejected by a human reviewer. Please rewrite the article incorporating their feedback.
Human Feedback: {humanResponse.Feedback}
""",
thread: writerThread);
content = writerResponse.Result;
}
// If we reach here, it means we exhausted the maximum number of iterations
throw new InvalidOperationException(
$"Content could not be approved after {input.MaxReviewAttempts} iterations.");
}
// Activity functions
static void NotifyUserForApproval(TaskActivityContext context, GeneratedContent content)
{
// In a real implementation, this would send notifications via email, SMS, etc.
Console.ForegroundColor = ConsoleColor.DarkMagenta;
Console.WriteLine(
$"""
NOTIFICATION: Please review the following content for approval:
Title: {content.Title}
Content: {content.Content}
""");
Console.ResetColor();
}
static void PublishContent(TaskActivityContext context, GeneratedContent content)
{
// In a real implementation, this would publish to a CMS, website, etc.
Console.ForegroundColor = ConsoleColor.DarkMagenta;
Console.WriteLine(
$"""
PUBLISHING: Content has been published successfully.
Title: {content.Title}
Content: {content.Content}
""");
Console.ResetColor();
}
// Tools that demonstrate starting orchestrations from agent tool calls.
[Description("Starts a content generation workflow and returns the instance ID for tracking.")]
static string StartContentGenerationWorkflow([Description("The topic for content generation")] string topic)
{
const int MaxReviewAttempts = 3;
const float ApprovalTimeoutHours = 72;
// Schedule the orchestration, which will start running after the tool call completes.
string instanceId = DurableAgentContext.Current.ScheduleNewOrchestration(
name: nameof(RunOrchestratorAsync),
input: new ContentGenerationInput
{
Topic = topic,
MaxReviewAttempts = MaxReviewAttempts,
ApprovalTimeoutHours = ApprovalTimeoutHours
});
return $"Workflow started with instance ID: {instanceId}";
}
[Description("Gets the status of a workflow orchestration and returns a summary of the workflow's current status.")]
static async Task<object> GetWorkflowStatusAsync(
[Description("The instance ID of the workflow to check")] string instanceId,
[Description("Whether to include detailed information")] bool includeDetails = true)
{
// Get the current agent context using the thread-static property
OrchestrationMetadata? status = await DurableAgentContext.Current.GetOrchestrationStatusAsync(
instanceId,
includeDetails);
if (status is null)
{
return new
{
instanceId,
error = $"Workflow instance '{instanceId}' not found.",
};
}
return new
{
instanceId = status.InstanceId,
createdAt = status.CreatedAt,
executionStatus = status.RuntimeStatus,
workflowStatus = status.SerializedCustomStatus,
lastUpdatedAt = status.LastUpdatedAt,
failureDetails = status.FailureDetails
};
}
[Description(
"Raises a feedback event for the content generation workflow. If approved, the workflow will be published. " +
"If rejected, the workflow will generate new content.")]
static async Task SubmitHumanFeedbackAsync(
[Description("The instance ID of the workflow to submit feedback for")] string instanceId,
[Description("Feedback to submit")] HumanFeedbackResponse feedback)
{
await DurableAgentContext.Current.RaiseOrchestrationEventAsync(instanceId, HumanFeedbackEventName, feedback);
}
// Configure the console app to host the AI agents.
IHost host = Host.CreateDefaultBuilder(args)
.ConfigureLogging(loggingBuilder => loggingBuilder.SetMinimumLevel(LogLevel.Warning))
.ConfigureServices(services =>
{
services.ConfigureDurableAgents(
options =>
{
// Add the writer agent used by the orchestration
options.AddAIAgent(writerAgent);
// Define the agent that can start orchestrations from tool calls
options.AddAIAgentFactory(PublisherAgentName, sp =>
{
return client.GetChatClient(deploymentName).AsAIAgent(
instructions: PublisherAgentInstructions,
name: PublisherAgentName,
services: sp,
tools: [
AIFunctionFactory.Create(StartContentGenerationWorkflow),
AIFunctionFactory.Create(GetWorkflowStatusAsync),
AIFunctionFactory.Create(SubmitHumanFeedbackAsync),
]);
});
},
workerBuilder: builder =>
{
builder.UseDurableTaskScheduler(dtsConnectionString);
builder.AddTasks(registry =>
{
registry.AddOrchestratorFunc<ContentGenerationInput>(nameof(RunOrchestratorAsync), RunOrchestratorAsync);
registry.AddActivityFunc<GeneratedContent>(nameof(NotifyUserForApproval), NotifyUserForApproval);
registry.AddActivityFunc<GeneratedContent>(nameof(PublishContent), PublishContent);
});
},
clientBuilder: builder => builder.UseDurableTaskScheduler(dtsConnectionString));
})
.Build();
await host.StartAsync();
// Get the agent proxy from services
IServiceProvider services = host.Services;
AIAgent? agentProxy = services.GetKeyedService<AIAgent>(PublisherAgentName);
if (agentProxy == null)
{
Console.ForegroundColor = ConsoleColor.Red;
Console.Error.WriteLine("Agent 'Publisher' not found.");
Console.ResetColor();
Environment.Exit(1);
return;
}
// Console colors for better UX
Console.ForegroundColor = ConsoleColor.Cyan;
Console.WriteLine("=== Long Running Tools Sample ===");
Console.ResetColor();
Console.WriteLine("Enter a topic for the Publisher agent to write about (or 'exit' to quit):");
Console.WriteLine();
// Create a thread for the conversation
AgentThread thread = await agentProxy.GetNewThreadAsync();
using CancellationTokenSource cts = new();
Console.CancelKeyPress += (sender, e) =>
{
e.Cancel = true;
cts.Cancel();
};
while (!cts.Token.IsCancellationRequested)
{
// Read input from stdin
Console.ForegroundColor = ConsoleColor.Yellow;
Console.Write("You: ");
Console.ResetColor();
string? input = Console.ReadLine();
if (string.IsNullOrWhiteSpace(input) || input.Equals("exit", StringComparison.OrdinalIgnoreCase))
{
break;
}
// Run the agent
Console.ForegroundColor = ConsoleColor.Green;
Console.Write("Publisher: ");
Console.ResetColor();
try
{
AgentResponse agentResponse = await agentProxy.RunAsync(
message: input,
thread: thread,
cancellationToken: cts.Token);
Console.WriteLine(agentResponse.Text);
Console.WriteLine();
}
catch (Exception ex)
{
Console.ForegroundColor = ConsoleColor.Red;
Console.Error.WriteLine($"Error: {ex.Message}");
Console.ResetColor();
Console.WriteLine();
}
Console.WriteLine("(Press Enter to prompt the Publisher agent again)");
_ = Console.ReadLine();
}
await host.StopAsync();
@@ -0,0 +1,90 @@
# Long Running Tools Sample
This sample demonstrates how to use the durable agents extension to create a console app with agents that have long running tools. This sample builds on the [05_AgentOrchestration_HITL](../05_AgentOrchestration_HITL) sample by adding a publisher agent that can start and manage content generation workflows. A key difference is that the publisher agent knows the IDs of the workflows it starts, so it can check the status of the workflows and approve or reject them without being explicitly given the context (instance IDs, etc).
## Key Concepts Demonstrated
The same key concepts as the [05_AgentOrchestration_HITL](../05_AgentOrchestration_HITL) sample are demonstrated, but with the following additional concepts:
- **Long running tools**: Using `DurableAgentContext.Current` to start orchestrations from tool calls
- **Multi-agent orchestration**: Agents can start and manage workflows that orchestrate other agents
- **Human-in-the-loop (with delegation)**: The agent acts as an intermediary between the human and the workflow. The human remains in the loop, but delegates to the agent to start the workflow and approve or reject the content.
## Environment Setup
See the [README.md](../README.md) file in the parent directory for more information on how to configure the environment, including how to install and run common sample dependencies.
## Running the Sample
With the environment setup, you can run the sample:
```bash
cd dotnet/samples/DurableAgents/ConsoleApps/06_LongRunningTools
dotnet run --framework net10.0
```
The app will prompt you for input. You can interact with the Publisher agent:
```text
=== Long Running Tools Sample ===
Enter a topic for the Publisher agent to write about (or 'exit' to quit):
You: Start a content generation workflow for the topic 'The Future of Artificial Intelligence'
Publisher: The content generation workflow for the topic "The Future of Artificial Intelligence" has been successfully started, and the instance ID is **6a04276e8d824d8d941e1dc4142cc254**. If you need any further assistance or updates on the workflow, feel free to ask!
```
Behind the scenes, the publisher agent will:
1. Start the content generation workflow via a tool call
2. The workflow will generate initial content using the Writer agent and wait for human approval, which will be visible in the terminal
Once the workflow is waiting for human approval, you can send approval or rejection by prompting the publisher agent accordingly.
> [!NOTE]
> You must press Enter after each message to continue the conversation. The sample is set up this way because the workflow is running in the background and may write to the console asynchronously.
To tell the agent to rewrite the content with feedback, you can prompt it to reject the content with feedback.
```text
You: Reject the content with feedback: The article needs more technical depth and better examples.
Publisher: The content has been successfully rejected with the feedback: "The article needs more technical depth and better examples." The workflow will now generate new content based on this feedback.
```
Once you're satisfied with the content, you can approve it for publishing.
```text
You: Approve the content
Publisher: The content has been successfully approved for publishing. If you need any more assistance or have further requests, feel free to let me know!
```
Once the workflow has completed, you can get the status by prompting the publisher agent to give you the status.
```text
You: Get the status of the workflow you previously started
Publisher: The status of the workflow with instance ID **6a04276e8d824d8d941e1dc4142cc254** is as follows:
- **Execution Status:** Completed
- **Created At:** December 22, 2025, 23:08:13 UTC
- **Last Updated At:** December 22, 2025, 23:09:59 UTC
- **Workflow Status:**
- Message: Content published successfully at December 22, 2025, 23:09:59 UTC
- Human Feedback: Approved
```
## Viewing Agent and Orchestration State
You can view the state of both the agent and the orchestrations it starts in the Durable Task Scheduler dashboard:
1. Open your browser and navigate to `http://localhost:8082`
2. In the dashboard, you can see:
- **Agents**: View the state of the Publisher agent, including its conversation history and tool call history
- **Orchestrations**: View the content generation orchestration instances that were started by the agent via tool calls, including their runtime status, custom status, input, output, and execution history
When the publisher agent starts a workflow, the orchestration instance ID is included in the agent's response. You can use this ID to find the specific orchestration in the dashboard and inspect:
- The orchestration's execution progress
- When it's waiting for human approval (visible in custom status)
- The content generation workflow state
- The WriterAgent state within the orchestration
This demonstrates how agents can manage long-running workflows and how you can monitor both the agent's state and the workflows it orchestrates.
@@ -0,0 +1,31 @@
<Project Sdk="Microsoft.NET.Sdk">
<PropertyGroup>
<TargetFrameworks>net10.0</TargetFrameworks>
<OutputType>Exe</OutputType>
<ImplicitUsings>enable</ImplicitUsings>
<Nullable>enable</Nullable>
<AssemblyName>ReliableStreaming</AssemblyName>
<RootNamespace>ReliableStreaming</RootNamespace>
</PropertyGroup>
<ItemGroup>
<PackageReference Include="Azure.AI.OpenAI" />
<PackageReference Include="Azure.Identity" />
<PackageReference Include="Microsoft.DurableTask.Client.AzureManaged" />
<PackageReference Include="Microsoft.DurableTask.Worker.AzureManaged" />
<PackageReference Include="Microsoft.Extensions.Hosting" />
<PackageReference Include="StackExchange.Redis" />
</ItemGroup>
<!-- Local projects that should be switched to package references when using the sample outside of this MAF repo -->
<!--
<ItemGroup>
<PackageReference Include="Microsoft.Agents.AI.DurableTask" />
<PackageReference Include="Microsoft.Agents.AI.OpenAI" />
</ItemGroup>
-->
<ItemGroup>
<ProjectReference Include="..\..\..\..\src\Microsoft.Agents.AI.DurableTask\Microsoft.Agents.AI.DurableTask.csproj" />
<ProjectReference Include="..\..\..\..\src\Microsoft.Agents.AI.OpenAI\Microsoft.Agents.AI.OpenAI.csproj" />
</ItemGroup>
</Project>
@@ -0,0 +1,363 @@
// Copyright (c) Microsoft. All rights reserved.
// This sample demonstrates how to implement reliable streaming for durable agents using Redis Streams.
// It reads prompts from stdin and streams agent responses to stdout in real-time.
using System.ComponentModel;
using Azure;
using Azure.AI.OpenAI;
using Azure.Identity;
using Microsoft.Agents.AI;
using Microsoft.Agents.AI.DurableTask;
using Microsoft.DurableTask.Client.AzureManaged;
using Microsoft.DurableTask.Worker.AzureManaged;
using Microsoft.Extensions.AI;
using Microsoft.Extensions.DependencyInjection;
using Microsoft.Extensions.Hosting;
using Microsoft.Extensions.Logging;
using OpenAI.Chat;
using ReliableStreaming;
using StackExchange.Redis;
// Get the Azure OpenAI endpoint and deployment name from environment variables.
string endpoint = Environment.GetEnvironmentVariable("AZURE_OPENAI_ENDPOINT")
?? throw new InvalidOperationException("AZURE_OPENAI_ENDPOINT is not set.");
string deploymentName = Environment.GetEnvironmentVariable("AZURE_OPENAI_DEPLOYMENT")
?? throw new InvalidOperationException("AZURE_OPENAI_DEPLOYMENT is not set.");
// Get Redis connection string from environment variable.
string redisConnectionString = Environment.GetEnvironmentVariable("REDIS_CONNECTION_STRING")
?? "localhost:6379";
// Get the Redis stream TTL from environment variable (default: 10 minutes).
int redisStreamTtlMinutes = int.Parse(Environment.GetEnvironmentVariable("REDIS_STREAM_TTL_MINUTES") ?? "10");
// Get DTS connection string from environment variable
string dtsConnectionString = Environment.GetEnvironmentVariable("DURABLE_TASK_SCHEDULER_CONNECTION_STRING")
?? "Endpoint=http://localhost:8080;TaskHub=default;Authentication=None";
// Use Azure Key Credential if provided, otherwise use Azure CLI Credential.
string? azureOpenAiKey = Environment.GetEnvironmentVariable("AZURE_OPENAI_KEY");
AzureOpenAIClient client = !string.IsNullOrEmpty(azureOpenAiKey)
? new AzureOpenAIClient(new Uri(endpoint), new AzureKeyCredential(azureOpenAiKey))
: new AzureOpenAIClient(new Uri(endpoint), new AzureCliCredential());
// Travel Planner agent instructions - designed to produce longer responses for demonstrating streaming.
const string TravelPlannerName = "TravelPlanner";
const string TravelPlannerInstructions =
"""
You are an expert travel planner who creates detailed, personalized travel itineraries.
When asked to plan a trip, you should:
1. Create a comprehensive day-by-day itinerary
2. Include specific recommendations for activities, restaurants, and attractions
3. Provide practical tips for each destination
4. Consider weather and local events when making recommendations
5. Include estimated times and logistics between activities
Always use the available tools to get current weather forecasts and local events
for the destination to make your recommendations more relevant and timely.
Format your response with clear headings for each day and include emoji icons
to make the itinerary easy to scan and visually appealing.
""";
// Mock travel tools that return hardcoded data for demonstration purposes.
[Description("Gets the weather forecast for a destination on a specific date. Use this to provide weather-aware recommendations in the itinerary.")]
static string GetWeatherForecast(string destination, string date)
{
Dictionary<string, (string condition, int highF, int lowF)> weatherByRegion = new(StringComparer.OrdinalIgnoreCase)
{
["Tokyo"] = ("Partly cloudy with a chance of light rain", 58, 45),
["Paris"] = ("Overcast with occasional drizzle", 52, 41),
["New York"] = ("Clear and cold", 42, 28),
["London"] = ("Foggy morning, clearing in afternoon", 48, 38),
["Sydney"] = ("Sunny and warm", 82, 68),
["Rome"] = ("Sunny with light breeze", 62, 48),
["Barcelona"] = ("Partly sunny", 59, 47),
["Amsterdam"] = ("Cloudy with light rain", 46, 38),
["Dubai"] = ("Sunny and hot", 85, 72),
["Singapore"] = ("Tropical thunderstorms in afternoon", 88, 77),
["Bangkok"] = ("Hot and humid, afternoon showers", 91, 78),
["Los Angeles"] = ("Sunny and pleasant", 72, 55),
["San Francisco"] = ("Morning fog, afternoon sun", 62, 52),
["Seattle"] = ("Rainy with breaks", 48, 40),
["Miami"] = ("Warm and sunny", 78, 65),
["Honolulu"] = ("Tropical paradise weather", 82, 72),
};
(string condition, int highF, int lowF) forecast = ("Partly cloudy", 65, 50);
foreach (KeyValuePair<string, (string, int, int)> entry in weatherByRegion)
{
if (destination.Contains(entry.Key, StringComparison.OrdinalIgnoreCase))
{
forecast = entry.Value;
break;
}
}
return $"""
Weather forecast for {destination} on {date}:
Conditions: {forecast.condition}
High: {forecast.highF}°F ({(forecast.highF - 32) * 5 / 9}°C)
Low: {forecast.lowF}°F ({(forecast.lowF - 32) * 5 / 9}°C)
Recommendation: {GetWeatherRecommendation(forecast.condition)}
""";
}
[Description("Gets local events and activities happening at a destination around a specific date. Use this to suggest timely activities and experiences.")]
static string GetLocalEvents(string destination, string date)
{
Dictionary<string, string[]> eventsByCity = new(StringComparer.OrdinalIgnoreCase)
{
["Tokyo"] = [
"🎭 Kabuki Theater Performance at Kabukiza Theatre - Traditional Japanese drama",
"🌸 Winter Illuminations at Yoyogi Park - Spectacular light displays",
"🍜 Ramen Festival at Tokyo Station - Sample ramen from across Japan",
"🎮 Gaming Expo at Tokyo Big Sight - Latest video games and technology",
],
["Paris"] = [
"🎨 Impressionist Exhibition at Musée d'Orsay - Extended evening hours",
"🍷 Wine Tasting Tour in Le Marais - Local sommelier guided",
"🎵 Jazz Night at Le Caveau de la Huchette - Historic jazz club",
"🥐 French Pastry Workshop - Learn from master pâtissiers",
],
["New York"] = [
"🎭 Broadway Show: Hamilton - Limited engagement performances",
"🏀 Knicks vs Lakers at Madison Square Garden",
"🎨 Modern Art Exhibit at MoMA - New installations",
"🍕 Pizza Walking Tour of Brooklyn - Artisan pizzerias",
],
["London"] = [
"👑 Royal Collection Exhibition at Buckingham Palace",
"🎭 West End Musical: The Phantom of the Opera",
"🍺 Craft Beer Festival at Brick Lane",
"🎪 Winter Wonderland at Hyde Park - Rides and markets",
],
["Sydney"] = [
"🏄 Pro Surfing Competition at Bondi Beach",
"🎵 Opera at Sydney Opera House - La Bohème",
"🦘 Wildlife Night Safari at Taronga Zoo",
"🍽️ Harbor Dinner Cruise with fireworks",
],
["Rome"] = [
"🏛️ After-Hours Vatican Tour - Skip the crowds",
"🍝 Pasta Making Class in Trastevere",
"🎵 Classical Concert at Borghese Gallery",
"🍷 Wine Tasting in Roman Cellars",
],
};
string[] events = [
"🎭 Local theater performance",
"🍽️ Food and wine festival",
"🎨 Art gallery opening",
"🎵 Live music at local venues",
];
foreach (KeyValuePair<string, string[]> entry in eventsByCity)
{
if (destination.Contains(entry.Key, StringComparison.OrdinalIgnoreCase))
{
events = entry.Value;
break;
}
}
string eventList = string.Join("\n• ", events);
return $"""
Local events in {destination} around {date}:
• {eventList}
💡 Tip: Book popular events in advance as they may sell out quickly!
""";
}
static string GetWeatherRecommendation(string condition)
{
return condition switch
{
string c when c.Contains("rain", StringComparison.OrdinalIgnoreCase) || c.Contains("drizzle", StringComparison.OrdinalIgnoreCase) =>
"Bring an umbrella and waterproof jacket. Consider indoor activities for backup.",
string c when c.Contains("fog", StringComparison.OrdinalIgnoreCase) =>
"Morning visibility may be limited. Plan outdoor sightseeing for afternoon.",
string c when c.Contains("cold", StringComparison.OrdinalIgnoreCase) =>
"Layer up with warm clothing. Hot drinks and cozy cafés recommended.",
string c when c.Contains("hot", StringComparison.OrdinalIgnoreCase) || c.Contains("warm", StringComparison.OrdinalIgnoreCase) =>
"Stay hydrated and use sunscreen. Plan strenuous activities for cooler morning hours.",
string c when c.Contains("thunder", StringComparison.OrdinalIgnoreCase) || c.Contains("storm", StringComparison.OrdinalIgnoreCase) =>
"Keep an eye on weather updates. Have indoor alternatives ready.",
_ => "Pleasant conditions expected. Great day for outdoor exploration!"
};
}
// Configure the console app to host the AI agent.
IHost host = Host.CreateDefaultBuilder(args)
.ConfigureLogging(loggingBuilder => loggingBuilder.SetMinimumLevel(LogLevel.Warning))
.ConfigureServices(services =>
{
services.ConfigureDurableAgents(
options =>
{
// Define the Travel Planner agent with tools for weather and events
options.AddAIAgentFactory(TravelPlannerName, sp =>
{
return client.GetChatClient(deploymentName).AsAIAgent(
instructions: TravelPlannerInstructions,
name: TravelPlannerName,
services: sp,
tools: [
AIFunctionFactory.Create(GetWeatherForecast),
AIFunctionFactory.Create(GetLocalEvents),
]);
});
},
workerBuilder: builder => builder.UseDurableTaskScheduler(dtsConnectionString),
clientBuilder: builder => builder.UseDurableTaskScheduler(dtsConnectionString));
// Register Redis connection as a singleton
services.AddSingleton<IConnectionMultiplexer>(_ =>
ConnectionMultiplexer.Connect(redisConnectionString));
// Register the Redis stream response handler - this captures agent responses
// and publishes them to Redis Streams for reliable delivery.
services.AddSingleton(sp =>
new RedisStreamResponseHandler(
sp.GetRequiredService<IConnectionMultiplexer>(),
TimeSpan.FromMinutes(redisStreamTtlMinutes)));
services.AddSingleton<IAgentResponseHandler>(sp =>
sp.GetRequiredService<RedisStreamResponseHandler>());
})
.Build();
await host.StartAsync();
// Get the agent proxy from services
IServiceProvider services = host.Services;
AIAgent? agentProxy = services.GetKeyedService<AIAgent>(TravelPlannerName);
RedisStreamResponseHandler streamHandler = services.GetRequiredService<RedisStreamResponseHandler>();
if (agentProxy == null)
{
Console.ForegroundColor = ConsoleColor.Red;
Console.Error.WriteLine($"Agent '{TravelPlannerName}' not found.");
Console.ResetColor();
Environment.Exit(1);
return;
}
// Console colors for better UX
Console.ForegroundColor = ConsoleColor.Cyan;
Console.WriteLine("=== Reliable Streaming Sample ===");
Console.ResetColor();
Console.WriteLine("Enter a travel planning request (or 'exit' to quit):");
Console.WriteLine();
string? lastCursor = null;
async Task ReadStreamTask(string conversationId, string? cursor, CancellationToken cancellationToken)
{
// Initialize lastCursor to the starting cursor position
// This ensures we have a valid cursor even if cancellation happens before any chunks are processed
lastCursor = cursor;
await foreach (StreamChunk chunk in streamHandler.ReadStreamAsync(conversationId, cursor, cancellationToken))
{
if (chunk.Error != null)
{
Console.ForegroundColor = ConsoleColor.Red;
Console.Error.WriteLine($"\n[Error: {chunk.Error}]");
Console.ResetColor();
break;
}
if (chunk.IsDone)
{
Console.WriteLine();
Console.WriteLine();
break;
}
if (chunk.Text != null)
{
Console.Write(chunk.Text);
}
// Always update lastCursor to track the latest entry ID, even if text is null
// This ensures we can resume from the correct position after interruption
if (!string.IsNullOrEmpty(chunk.EntryId))
{
lastCursor = chunk.EntryId;
}
}
}
// New conversation: prompt from stdin
Console.ForegroundColor = ConsoleColor.Yellow;
Console.Write("You: ");
Console.ResetColor();
string? prompt = Console.ReadLine();
if (string.IsNullOrWhiteSpace(prompt) || prompt.Equals("exit", StringComparison.OrdinalIgnoreCase))
{
return;
}
// Create a new agent thread
AgentThread thread = await agentProxy.GetNewThreadAsync();
AgentSessionId sessionId = thread.GetService<AgentSessionId>();
string conversationId = sessionId.ToString();
Console.ForegroundColor = ConsoleColor.Green;
Console.WriteLine($"Conversation ID: {conversationId}");
Console.WriteLine("Press [Enter] to interrupt the stream.");
Console.ResetColor();
// Run the agent in the background
DurableAgentRunOptions options = new() { IsFireAndForget = true };
await agentProxy.RunAsync(prompt, thread, options, CancellationToken.None);
bool streamCompleted = false;
while (!streamCompleted)
{
// On a key press, cancel the cancellation token to stop the stream
using CancellationTokenSource userCancellationSource = new();
_ = Task.Run(() =>
{
_ = Console.ReadLine();
userCancellationSource.Cancel();
});
try
{
// Start reading the stream and wait for it to complete
await ReadStreamTask(conversationId, lastCursor, userCancellationSource.Token);
streamCompleted = true;
}
catch (OperationCanceledException)
{
Console.ForegroundColor = ConsoleColor.Yellow;
Console.WriteLine("Stream cancelled. Press [Enter] to reconnect and resume the stream from the last cursor.");
// Ensure lastCursor is set - if it's still null, we at least have the starting cursor
string cursorValue = lastCursor ?? "(n/a)";
Console.WriteLine($"Last cursor: {cursorValue}");
Console.ResetColor();
// Explicitly flush to ensure the message is written immediately
Console.Out.Flush();
}
if (!streamCompleted)
{
Console.ReadLine();
Console.ForegroundColor = ConsoleColor.Green;
Console.WriteLine($"Resuming conversation: {conversationId} from cursor: {lastCursor ?? "(beginning)"}");
Console.ResetColor();
}
}
Console.ForegroundColor = ConsoleColor.Green;
Console.WriteLine("Conversation completed.");
Console.ResetColor();
await host.StopAsync();
@@ -0,0 +1,181 @@
# Reliable Streaming with Redis
This sample demonstrates how to implement reliable streaming for durable agents using Redis Streams as a message broker. It enables clients to disconnect and reconnect to ongoing agent responses without losing messages, inspired by [OpenAI's background mode](https://platform.openai.com/docs/guides/background) for the Responses API.
## Key Concepts Demonstrated
- **Reliable message delivery**: Agent responses are persisted to Redis Streams, allowing clients to resume from any point
- **Real-time streaming**: Chunks are printed to stdout as they arrive (like `tail -f`)
- **Cursor-based resumption**: Each chunk includes an entry ID that can be used to resume the stream
- **Fire-and-forget agent invocation**: The agent runs in the background while the client streams from Redis
## Environment Setup
See the [README.md](../README.md) file in the parent directory for more information on how to configure the environment, including how to install and run common sample dependencies.
### Additional Requirements: Redis
This sample requires a Redis instance. Start a local Redis instance using Docker:
```bash
docker run -d --name redis -p 6379:6379 redis:latest
```
To verify Redis is running:
```bash
docker ps | grep redis
```
## Running the Sample
With the environment setup, you can run the sample:
```bash
cd dotnet/samples/DurableAgents/ConsoleApps/07_ReliableStreaming
dotnet run --framework net10.0
```
The app will prompt you for a travel planning request:
```text
=== Reliable Streaming Sample ===
Enter a travel planning request (or 'exit' to quit):
You: Plan a 7-day trip to Tokyo, Japan for next month. Include daily activities, restaurant recommendations, and tips for getting around.
```
The agent's response will stream to your console in real-time as chunks arrive from Redis:
```text
Starting new conversation: @dafx-travelplanner@a1b2c3d4e5f67890abcdef1234567890
Press [Enter] to interrupt the stream.
TravelPlanner: # 7-Day Tokyo Adventure
## Day 1: Arrival and Exploration
...
```
### Demonstrating Stream Interruption and Resumption
This is the key feature of reliable streaming. Follow these steps to see it in action:
1. **Start a stream**: Run the app and enter a travel planning request
2. **Note the conversation ID**: The conversation ID is displayed at the start of the stream (e.g., `Starting new conversation: @dafx-travelplanner@a1b2c3d4e5f67890abcdef1234567890`)
3. **Interrupt the stream**: While the agent is still generating text, press **`Enter`** to interrupt. The agent continues running in the background - your messages are being saved to Redis.
4. **Resume the stream**: Press **`Enter`** again to reconnect and resume the stream from the last cursor position. The app will automatically resume from where it left off.
```text
Starting new conversation: @dafx-travelplanner@a1b2c3d4e5f67890abcdef1234567890
Press [Enter] to interrupt the stream.
TravelPlanner: # 7-Day Tokyo Adventure
## Day 1: Arrival and Exploration
[Streaming content...]
[Press Enter to interrupt]
Stream cancelled. Press [Enter] to reconnect and resume the stream from the last cursor.
Last cursor: 1734567890123-0
[Press Enter to resume]
Resuming conversation: @dafx-travelplanner@a1b2c3d4e5f67890abcdef1234567890 from cursor: 1734567890123-0
[Stream continues from where it left off...]
```
## Viewing Agent State
You can view the state of the agent in the Durable Task Scheduler dashboard:
1. Open your browser and navigate to `http://localhost:8082`
2. In the dashboard, you can see:
- **Agents**: View the state of the TravelPlanner agent, including conversation history and current state
- **Orchestrations**: View any orchestrations that may have been triggered by the agent
The conversation ID displayed in the console output (shown as "Starting new conversation: {conversationId}") corresponds to the agent's conversation thread. You can use this to identify the agent in the dashboard and inspect:
- The agent's conversation state
- Tool calls made by the agent (weather and events lookups)
- The streaming response state
Note that while the console app streams responses from Redis, the agent state in DTS shows the underlying durable agent execution, including all tool calls and conversation context.
## Architecture Overview
```text
┌─────────────┐ stdin (prompt) ┌─────────────────────┐
│ Client │ ─────────────────────► │ Console App │
│ (stdin) │ │ (Program.cs) │
└─────────────┘ └──────────────┬──────┘
▲ │
│ stdout (chunks) Signal Entity
│ │
│ ▼
│ ┌─────────────────────┐
│ │ AgentEntity │
│ │ (Durable Entity) │
│ └──────────┬──────────┘
│ │
│ IAgentResponseHandler
│ │
│ ▼
│ ┌─────────────────────┐
│ │ RedisStreamResponse │
│ │ Handler │
│ └──────────┬──────────┘
│ │
│ XADD (write)
│ │
│ ▼
│ ┌─────────────────────┐
└─────────── XREAD (poll) ────────── │ Redis Streams │
│ (Durable Log) │
└─────────────────────┘
```
### Data Flow
1. **Client sends prompt**: The console app reads the prompt from stdin and generates a new agent thread.
2. **Agent invoked**: The durable agent is signaled to run the travel planner agent. This is fire-and-forget from the console app's perspective.
3. **Responses captured**: As the agent generates responses, the `RedisStreamResponseHandler` (implementing `IAgentResponseHandler`) extracts the text from each `AgentRunResponseUpdate` and publishes it to a Redis Stream keyed by the agent session's conversation ID.
4. **Client polls Redis**: The console app streams events by polling the Redis Stream and printing chunks to stdout as they arrive.
5. **Resumption**: If the client interrupts the stream (e.g., by pressing Enter in the sample), it can resume from the last cursor position by providing the conversation ID and cursor to the call to resume the stream.
## Message Delivery Guarantees
This sample provides **at-least-once delivery** with the following characteristics:
- **Durability**: Messages are persisted to Redis Streams with configurable TTL (default: 10 minutes).
- **Ordering**: Messages are delivered in order within a session.
- **Real-time**: Chunks are printed as soon as they arrive from Redis.
### Important Considerations
- **No exactly-once delivery**: If a client disconnects exactly when receiving a message, it may receive that message again upon resumption. Clients should handle duplicate messages idempotently.
- **TTL expiration**: Streams expire after the configured TTL. Clients cannot resume streams that have expired.
- **Redis guarantees**: Redis streams are backed by Redis persistence mechanisms (RDB/AOF). Ensure your Redis instance is configured for durability as needed.
## Configuration
| Environment Variable | Description | Default |
|---------------------|-------------|---------|
| `REDIS_CONNECTION_STRING` | Redis connection string | `localhost:6379` |
| `REDIS_STREAM_TTL_MINUTES` | How long streams are retained after last write | `10` |
| `AZURE_OPENAI_ENDPOINT` | Azure OpenAI endpoint URL | (required) |
| `AZURE_OPENAI_DEPLOYMENT` | Azure OpenAI deployment name | (required) |
| `AZURE_OPENAI_KEY` | API key (optional, uses Azure CLI auth if not set) | (optional) |
## Cleanup
To stop and remove the Redis Docker containers:
```bash
docker stop redis
docker rm redis
```
@@ -0,0 +1,216 @@
// Copyright (c) Microsoft. All rights reserved.
using System.Runtime.CompilerServices;
using Microsoft.Agents.AI;
using Microsoft.Agents.AI.DurableTask;
using StackExchange.Redis;
namespace ReliableStreaming;
/// <summary>
/// Represents a chunk of data read from a Redis stream.
/// </summary>
/// <param name="EntryId">The Redis stream entry ID (can be used as a cursor for resumption).</param>
/// <param name="Text">The text content of the chunk, or null if this is a completion/error marker.</param>
/// <param name="IsDone">True if this chunk marks the end of the stream.</param>
/// <param name="Error">An error message if something went wrong, or null otherwise.</param>
public readonly record struct StreamChunk(string EntryId, string? Text, bool IsDone, string? Error);
/// <summary>
/// An implementation of <see cref="IAgentResponseHandler"/> that publishes agent response updates
/// to Redis Streams for reliable delivery. This enables clients to disconnect and reconnect
/// to ongoing agent responses without losing messages.
/// </summary>
/// <remarks>
/// <para>
/// Redis Streams provide a durable, append-only log that supports consumer groups and message
/// acknowledgment. This implementation uses auto-generated IDs (which are timestamp-based)
/// as sequence numbers, allowing clients to resume from any point in the stream.
/// </para>
/// <para>
/// Each agent session gets its own Redis Stream, keyed by session ID. The stream entries
/// contain text chunks extracted from <see cref="AgentResponseUpdate"/> objects.
/// </para>
/// </remarks>
public sealed class RedisStreamResponseHandler : IAgentResponseHandler
{
private const int MaxEmptyReads = 300; // 5 minutes at 1 second intervals
private const int PollIntervalMs = 1000;
private readonly IConnectionMultiplexer _redis;
private readonly TimeSpan _streamTtl;
/// <summary>
/// Initializes a new instance of the <see cref="RedisStreamResponseHandler" /> class.
/// </summary>
/// <param name="redis">The Redis connection multiplexer.</param>
/// <param name="streamTtl">The time-to-live for stream entries. Streams will expire after this duration of inactivity.</param>
public RedisStreamResponseHandler(IConnectionMultiplexer redis, TimeSpan streamTtl)
{
this._redis = redis;
this._streamTtl = streamTtl;
}
/// <inheritdoc/>
public async ValueTask OnStreamingResponseUpdateAsync(
IAsyncEnumerable<AgentResponseUpdate> messageStream,
CancellationToken cancellationToken)
{
// Get the current session ID from the DurableAgentContext
// This is set by the AgentEntity before invoking the response handler
DurableAgentContext context = DurableAgentContext.Current
?? throw new InvalidOperationException("DurableAgentContext.Current is not set. This handler must be used within a durable agent context.");
// Get conversation ID from the current thread context, which is only available in the context of
// a durable agent execution.
string conversationId = context.CurrentThread.GetService<AgentSessionId>().ToString();
if (string.IsNullOrEmpty(conversationId))
{
throw new InvalidOperationException("Unable to determine conversation ID from the current thread.");
}
string streamKey = GetStreamKey(conversationId);
IDatabase db = this._redis.GetDatabase();
int sequenceNumber = 0;
await foreach (AgentResponseUpdate update in messageStream.WithCancellation(cancellationToken))
{
// Extract just the text content - this avoids serialization round-trip issues
string text = update.Text;
// Only publish non-empty text chunks
if (!string.IsNullOrEmpty(text))
{
// Create the stream entry with the text and metadata
NameValueEntry[] entries =
[
new NameValueEntry("text", text),
new NameValueEntry("sequence", sequenceNumber++),
new NameValueEntry("timestamp", DateTimeOffset.UtcNow.ToUnixTimeMilliseconds()),
];
// Add to the Redis Stream with auto-generated ID (timestamp-based)
await db.StreamAddAsync(streamKey, entries);
// Refresh the TTL on each write to keep the stream alive during active streaming
await db.KeyExpireAsync(streamKey, this._streamTtl);
}
}
// Add a sentinel entry to mark the end of the stream
NameValueEntry[] endEntries =
[
new NameValueEntry("text", ""),
new NameValueEntry("sequence", sequenceNumber),
new NameValueEntry("timestamp", DateTimeOffset.UtcNow.ToUnixTimeMilliseconds()),
new NameValueEntry("done", "true"),
];
await db.StreamAddAsync(streamKey, endEntries);
// Set final TTL - the stream will be cleaned up after this duration
await db.KeyExpireAsync(streamKey, this._streamTtl);
}
/// <inheritdoc/>
public ValueTask OnAgentResponseAsync(AgentResponse message, CancellationToken cancellationToken)
{
// This handler is optimized for streaming responses.
// For non-streaming responses, we don't need to store in Redis since
// the response is returned directly to the caller.
return ValueTask.CompletedTask;
}
/// <summary>
/// Reads chunks from a Redis stream for the given session, yielding them as they become available.
/// </summary>
/// <param name="conversationId">The conversation ID to read from.</param>
/// <param name="cursor">Optional cursor to resume from. If null, reads from the beginning.</param>
/// <param name="cancellationToken">Cancellation token.</param>
/// <returns>An async enumerable of stream chunks.</returns>
public async IAsyncEnumerable<StreamChunk> ReadStreamAsync(
string conversationId,
string? cursor,
[EnumeratorCancellation] CancellationToken cancellationToken)
{
string streamKey = GetStreamKey(conversationId);
IDatabase db = this._redis.GetDatabase();
string startId = string.IsNullOrEmpty(cursor) ? "0-0" : cursor;
int emptyReadCount = 0;
bool hasSeenData = false;
while (!cancellationToken.IsCancellationRequested)
{
StreamEntry[]? entries = null;
string? errorMessage = null;
try
{
entries = await db.StreamReadAsync(streamKey, startId, count: 100);
}
catch (Exception ex)
{
errorMessage = ex.Message;
}
if (errorMessage != null)
{
yield return new StreamChunk(startId, null, false, errorMessage);
yield break;
}
// entries is guaranteed to be non-null if errorMessage is null
if (entries!.Length == 0)
{
if (!hasSeenData)
{
emptyReadCount++;
if (emptyReadCount >= MaxEmptyReads)
{
yield return new StreamChunk(
startId,
null,
false,
$"Stream not found or timed out after {MaxEmptyReads * PollIntervalMs / 1000} seconds");
yield break;
}
}
await Task.Delay(PollIntervalMs, cancellationToken);
continue;
}
hasSeenData = true;
foreach (StreamEntry entry in entries)
{
startId = entry.Id.ToString();
string? text = entry["text"];
string? done = entry["done"];
if (done == "true")
{
yield return new StreamChunk(startId, null, true, null);
yield break;
}
if (!string.IsNullOrEmpty(text))
{
yield return new StreamChunk(startId, text, false, null);
}
}
}
// If we exited the loop due to cancellation, throw to signal the caller
cancellationToken.ThrowIfCancellationRequested();
}
/// <summary>
/// Gets the Redis Stream key for a given conversation ID.
/// </summary>
/// <param name="conversationId">The conversation ID.</param>
/// <returns>The Redis Stream key.</returns>
internal static string GetStreamKey(string conversationId) => $"agent-stream:{conversationId}";
}
@@ -0,0 +1,109 @@
# Console App Samples
This directory contains samples for console app hosting of durable agents. These samples use standard I/O (stdin/stdout) for interaction, making them both interactive and scriptable.
- **[01_SingleAgent](01_SingleAgent)**: A sample that demonstrates how to host a single conversational agent in a console app and interact with it via stdin/stdout.
- **[02_AgentOrchestration_Chaining](02_AgentOrchestration_Chaining)**: A sample that demonstrates how to host a single conversational agent in a console app and invoke it using a durable orchestration.
- **[03_AgentOrchestration_Concurrency](03_AgentOrchestration_Concurrency)**: A sample that demonstrates how to host multiple agents in a console app and run them concurrently using a durable orchestration.
- **[04_AgentOrchestration_Conditionals](04_AgentOrchestration_Conditionals)**: A sample that demonstrates how to host multiple agents in a console app and run them sequentially using a durable orchestration with conditionals.
- **[05_AgentOrchestration_HITL](05_AgentOrchestration_HITL)**: A sample that demonstrates how to implement a human-in-the-loop workflow using durable orchestration, including interactive approval prompts.
- **[06_LongRunningTools](06_LongRunningTools)**: A sample that demonstrates how agents can start and interact with durable orchestrations from tool calls to enable long-running tool scenarios.
- **[07_ReliableStreaming](07_ReliableStreaming)**: A sample that demonstrates how to implement reliable streaming for durable agents using Redis Streams, enabling clients to disconnect and reconnect without losing messages.
## Running the Samples
These samples are designed to be run locally in a cloned repository.
### Prerequisites
The following prerequisites are required to run the samples:
- [.NET 10.0 SDK or later](https://dotnet.microsoft.com/download/dotnet)
- [Azure CLI](https://learn.microsoft.com/cli/azure/install-azure-cli) installed and authenticated (`az login`) or an API key for the Azure OpenAI service
- [Azure OpenAI Service](https://learn.microsoft.com/azure/ai-services/openai/how-to/create-resource) with a deployed model (gpt-4o-mini or better is recommended)
- [Durable Task Scheduler](https://learn.microsoft.com/azure/azure-functions/durable/durable-task-scheduler/develop-with-durable-task-scheduler) (local emulator or Azure-hosted)
- [Docker](https://docs.docker.com/get-docker/) installed if running the Durable Task Scheduler emulator locally
- [Redis](https://redis.io/) (for sample 07 only) - can be run locally using Docker
### Configuring RBAC Permissions for Azure OpenAI
These samples are configured to use the Azure OpenAI service with RBAC permissions to access the model. You'll need to configure the RBAC permissions for the Azure OpenAI service to allow the console app to access the model.
Below is an example of how to configure the RBAC permissions for the Azure OpenAI service to allow the current user to access the model.
Bash (Linux/macOS/WSL):
```bash
az role assignment create \
--assignee "yourname@contoso.com" \
--role "Cognitive Services OpenAI User" \
--scope /subscriptions/<your-subscription-id>/resourceGroups/<your-resource-group-name>/providers/Microsoft.CognitiveServices/accounts/<your-openai-resource-name>
```
PowerShell:
```powershell
az role assignment create `
--assignee "yourname@contoso.com" `
--role "Cognitive Services OpenAI User" `
--scope /subscriptions/<your-subscription-id>/resourceGroups/<your-resource-group-name>/providers/Microsoft.CognitiveServices/accounts/<your-openai-resource-name>
```
More information on how to configure RBAC permissions for Azure OpenAI can be found in the [Azure OpenAI documentation](https://learn.microsoft.com/azure/ai-services/openai/how-to/create-resource?pivots=cli).
### Setting an API key for the Azure OpenAI service
As an alternative to configuring Azure RBAC permissions, you can set an API key for the Azure OpenAI service by setting the `AZURE_OPENAI_KEY` environment variable.
Bash (Linux/macOS/WSL):
```bash
export AZURE_OPENAI_KEY="your-api-key"
```
PowerShell:
```powershell
$env:AZURE_OPENAI_KEY="your-api-key"
```
### Start Durable Task Scheduler
Most samples use the Durable Task Scheduler (DTS) to support hosted agents and durable orchestrations. DTS also allows you to view the status of orchestrations and their inputs and outputs from a web UI.
To run the Durable Task Scheduler locally, you can use the following `docker` command:
```bash
docker run -d --name dts-emulator -p 8080:8080 -p 8082:8082 mcr.microsoft.com/dts/dts-emulator:latest
```
The DTS dashboard will be available at `http://localhost:8080`.
### Environment Configuration
Each sample reads configuration from environment variables. You'll need to set the following environment variables:
```bash
export AZURE_OPENAI_ENDPOINT="https://your-resource.openai.azure.com/"
export AZURE_OPENAI_DEPLOYMENT="your-deployment-name"
```
### Running the Console Apps
Navigate to the sample directory and run the console app:
```bash
cd dotnet/samples/DurableAgents/ConsoleApps/01_SingleAgent
dotnet run --framework net10.0
```
> [!NOTE]
> The `--framework` option is required to specify the target framework for the console app because the samples are designed to support multiple target frameworks. If you are using a different target framework, you can specify it with the `--framework` option.
The app will prompt you for input via stdin.
### Viewing the sample output
The console app output is displayed directly in the terminal where you ran `dotnet run`. Agent responses are printed to stdout with subtle color coding for better readability.
You can also see the state of agents and orchestrations in the Durable Task Scheduler dashboard at `http://localhost:8082`.
@@ -0,0 +1,9 @@
<Project>
<Import Project="../Directory.Build.props" />
<!-- Remove the Environment alias from parent Directory.Build.props to allow System.Environment usage -->
<ItemGroup>
<Using Remove="SampleHelpers.SampleEnvironment" />
</ItemGroup>
</Project>
@@ -23,14 +23,14 @@ A2ACardResolver agentCardResolver = new(new Uri(a2aAgentHost));
AgentCard agentCard = await agentCardResolver.GetAgentCardAsync();
// Create an instance of the AIAgent for an existing A2A agent specified by the agent card.
AIAgent a2aAgent = agentCard.GetAIAgent();
AIAgent a2aAgent = agentCard.AsAIAgent();
// Create the main agent, and provide the a2a agent skills as a function tools.
AIAgent agent = new AzureOpenAIClient(
new Uri(endpoint),
new AzureCliCredential())
.GetChatClient(deploymentName)
.CreateAIAgent(
.AsAIAgent(
instructions: "You are a helpful assistant that helps people with travel planning.",
tools: [.. CreateFunctionTools(a2aAgent, agentCard)]
);
@@ -14,7 +14,7 @@ A2ACardResolver agentCardResolver = new(new Uri(a2aAgentHost));
AgentCard agentCard = await agentCardResolver.GetAgentCardAsync();
// Create an instance of the AIAgent for an existing A2A agent specified by the agent card.
AIAgent agent = agentCard.GetAIAgent();
AIAgent agent = agentCard.AsAIAgent();
AgentThread thread = await agent.GetNewThreadAsync();
@@ -16,7 +16,7 @@ using HttpClient httpClient = new()
AGUIChatClient chatClient = new(httpClient, serverUrl);
AIAgent agent = chatClient.CreateAIAgent(
AIAgent agent = chatClient.AsAIAgent(
name: "agui-client",
description: "AG-UI Client Agent");
@@ -24,7 +24,7 @@ ChatClient chatClient = new AzureOpenAIClient(
new DefaultAzureCredential())
.GetChatClient(deploymentName);
AIAgent agent = chatClient.AsIChatClient().CreateAIAgent(
AIAgent agent = chatClient.AsIChatClient().AsAIAgent(
name: "AGUIAssistant",
instructions: "You are a helpful assistant.");
@@ -16,7 +16,7 @@ using HttpClient httpClient = new()
AGUIChatClient chatClient = new(httpClient, serverUrl);
AIAgent agent = chatClient.CreateAIAgent(
AIAgent agent = chatClient.AsAIAgent(
name: "agui-client",
description: "AG-UI Client Agent");
@@ -79,7 +79,7 @@ ChatClient chatClient = new AzureOpenAIClient(
new DefaultAzureCredential())
.GetChatClient(deploymentName);
ChatClientAgent agent = chatClient.AsIChatClient().CreateAIAgent(
ChatClientAgent agent = chatClient.AsIChatClient().AsAIAgent(
name: "AGUIAssistant",
instructions: "You are a helpful assistant with access to restaurant information.",
tools: tools);
@@ -28,7 +28,7 @@ using HttpClient httpClient = new()
AGUIChatClient chatClient = new(httpClient, serverUrl);
AIAgent agent = chatClient.CreateAIAgent(
AIAgent agent = chatClient.AsAIAgent(
name: "agui-client",
description: "AG-UI Client Agent",
tools: frontendTools);
@@ -24,7 +24,7 @@ ChatClient chatClient = new AzureOpenAIClient(
new DefaultAzureCredential())
.GetChatClient(deploymentName);
AIAgent agent = chatClient.AsIChatClient().CreateAIAgent(
AIAgent agent = chatClient.AsIChatClient().AsAIAgent(
name: "AGUIAssistant",
instructions: "You are a helpful assistant.");
@@ -16,7 +16,7 @@ using HttpClient httpClient = new()
AGUIChatClient chatClient = new(httpClient, serverUrl);
// Create agent
ChatClientAgent baseAgent = chatClient.CreateAIAgent(
ChatClientAgent baseAgent = chatClient.AsAIAgent(
name: "AGUIAssistant",
instructions: "You are a helpful assistant.");
@@ -57,7 +57,7 @@ ChatClient openAIChatClient = new AzureOpenAIClient(
new DefaultAzureCredential())
.GetChatClient(deploymentName);
ChatClientAgent baseAgent = openAIChatClient.AsIChatClient().CreateAIAgent(
ChatClientAgent baseAgent = openAIChatClient.AsIChatClient().AsAIAgent(
name: "AGUIAssistant",
instructions: "You are a helpful assistant in charge of approving expenses",
tools: tools);
@@ -19,7 +19,7 @@ using HttpClient httpClient = new()
AGUIChatClient chatClient = new(httpClient, serverUrl);
AIAgent baseAgent = chatClient.CreateAIAgent(
AIAgent baseAgent = chatClient.AsAIAgent(
name: "recipe-client",
description: "AG-UI Recipe Client Agent");
@@ -34,7 +34,7 @@ ChatClient chatClient = new AzureOpenAIClient(
new DefaultAzureCredential())
.GetChatClient(deploymentName);
AIAgent baseAgent = chatClient.AsIChatClient().CreateAIAgent(
AIAgent baseAgent = chatClient.AsIChatClient().AsAIAgent(
name: "RecipeAgent",
instructions: """
You are a helpful recipe assistant. When users ask you to create or suggest a recipe,
@@ -26,7 +26,7 @@ using Microsoft.Agents.AI.A2A;
A2AClient a2aClient = new(new Uri("https://your-a2a-agent-host/echo"));
// Create an AIAgent from the A2AClient
AIAgent agent = a2aClient.GetAIAgent();
AIAgent agent = a2aClient.AsAIAgent();
// Run the agent
AgentResponse response = await agent.RunAsync("Tell me a joke about a pirate.");
@@ -26,7 +26,7 @@ AnthropicClient? client = (resource is null)
? new AnthropicFoundryClient(new AnthropicFoundryApiKeyCredentials(apiKey, resource)) // If an apiKey is provided, use Foundry with ApiKey authentication
: new AnthropicFoundryClient(new AnthropicAzureTokenCredential(new AzureCliCredential(), resource)); // Otherwise, use Foundry with Azure Client authentication
AIAgent agent = client.CreateAIAgent(model: deploymentName, instructions: JokerInstructions, name: JokerName);
AIAgent agent = client.AsAIAgent(model: deploymentName, instructions: JokerInstructions, name: JokerName);
// Invoke the agent and output the text result.
Console.WriteLine(await agent.RunAsync("Tell me a joke about a pirate."));
@@ -26,14 +26,14 @@ var createdAgentVersion = aiProjectClient.Agents.CreateAgentVersion(agentName: J
// agentVersion.Version = <versionNumber>,
// agentVersion.Name = <agentName>
// You can retrieve an AIAgent for an already created server side agent version.
AIAgent existingJokerAgent = aiProjectClient.GetAIAgent(createdAgentVersion);
// You can use an AIAgent with an already created server side agent version.
AIAgent existingJokerAgent = aiProjectClient.AsAIAgent(createdAgentVersion);
// You can also create another AIAgent version by providing the same name with a different definition.
AIAgent newJokerAgent = aiProjectClient.CreateAIAgent(name: JokerName, model: deploymentName, instructions: "You are extremely hilarious at telling jokes.");
AIAgent newJokerAgent = await aiProjectClient.CreateAIAgentAsync(name: JokerName, model: deploymentName, instructions: "You are extremely hilarious at telling jokes.");
// You can also get the AIAgent latest version just providing its name.
AIAgent jokerAgentLatest = aiProjectClient.GetAIAgent(name: JokerName);
AIAgent jokerAgentLatest = await aiProjectClient.GetAIAgentAsync(name: JokerName);
var latestAgentVersion = jokerAgentLatest.GetService<AgentVersion>()!;
// The AIAgent version can be accessed via the GetService method.
@@ -25,7 +25,7 @@ OpenAIClient client = string.IsNullOrWhiteSpace(apiKey)
AIAgent agent = client
.GetChatClient(model)
.CreateAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
.AsAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
// Invoke the agent and output the text result.
Console.WriteLine(await agent.RunAsync("Tell me a joke about a pirate."));
@@ -14,7 +14,7 @@ AIAgent agent = new AzureOpenAIClient(
new Uri(endpoint),
new AzureCliCredential())
.GetChatClient(deploymentName)
.CreateAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
.AsAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
// Invoke the agent and output the text result.
Console.WriteLine(await agent.RunAsync("Tell me a joke about a pirate."));
@@ -14,7 +14,7 @@ AIAgent agent = new AzureOpenAIClient(
new Uri(endpoint),
new AzureCliCredential())
.GetResponsesClient(deploymentName)
.CreateAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
.AsAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
// Invoke the agent and output the text result.
Console.WriteLine(await agent.RunAsync("Tell me a joke about a pirate."));
@@ -1,558 +0,0 @@
// Copyright (c) Microsoft. All rights reserved.
using System.Runtime.CompilerServices;
using Google.Apis.Util;
using Google.GenAI;
using Google.GenAI.Types;
namespace Microsoft.Extensions.AI;
/// <summary>Provides an <see cref="IChatClient"/> implementation based on <see cref="Client"/>.</summary>
internal sealed class GoogleGenAIChatClient : IChatClient
{
/// <summary>The wrapped <see cref="Client"/> instance (optional).</summary>
private readonly Client? _client;
/// <summary>The wrapped <see cref="Models"/> instance.</summary>
private readonly Models _models;
/// <summary>The default model that should be used when no override is specified.</summary>
private readonly string? _defaultModelId;
/// <summary>Lazily-initialized metadata describing the implementation.</summary>
private ChatClientMetadata? _metadata;
/// <summary>Initializes a new <see cref="GoogleGenAIChatClient"/> instance.</summary>
public GoogleGenAIChatClient(Client client, string? defaultModelId)
{
this._client = client;
this._models = client.Models;
this._defaultModelId = defaultModelId;
}
/// <summary>Initializes a new <see cref="GoogleGenAIChatClient"/> instance.</summary>
public GoogleGenAIChatClient(Models client, string? defaultModelId)
{
this._models = client;
this._defaultModelId = defaultModelId;
}
/// <inheritdoc />
public async Task<ChatResponse> GetResponseAsync(IEnumerable<ChatMessage> messages, ChatOptions? options = null, CancellationToken cancellationToken = default)
{
Utilities.ThrowIfNull(messages, nameof(messages));
// Create the request.
(string? modelId, List<Content> contents, GenerateContentConfig config) = this.CreateRequest(messages, options);
// Send it.
GenerateContentResponse generateResult = await this._models.GenerateContentAsync(modelId!, contents, config).ConfigureAwait(false);
// Create the response.
ChatResponse chatResponse = new(new ChatMessage(ChatRole.Assistant, []))
{
CreatedAt = generateResult.CreateTime is { } dt ? new DateTimeOffset(dt) : null,
ModelId = !string.IsNullOrWhiteSpace(generateResult.ModelVersion) ? generateResult.ModelVersion : modelId,
RawRepresentation = generateResult,
ResponseId = generateResult.ResponseId,
};
// Populate the response messages.
chatResponse.FinishReason = PopulateResponseContents(generateResult, chatResponse.Messages[0].Contents);
// Populate usage information if there is any.
if (generateResult.UsageMetadata is { } usageMetadata)
{
chatResponse.Usage = ExtractUsageDetails(usageMetadata);
}
// Return the response.
return chatResponse;
}
/// <inheritdoc />
public async IAsyncEnumerable<ChatResponseUpdate> GetStreamingResponseAsync(IEnumerable<ChatMessage> messages, ChatOptions? options = null, [EnumeratorCancellation] CancellationToken cancellationToken = default)
{
Utilities.ThrowIfNull(messages, nameof(messages));
// Create the request.
(string? modelId, List<Content> contents, GenerateContentConfig config) = this.CreateRequest(messages, options);
// Send it, and process the results.
await foreach (GenerateContentResponse generateResult in this._models.GenerateContentStreamAsync(modelId!, contents, config).WithCancellation(cancellationToken).ConfigureAwait(false))
{
// Create a response update for each result in the stream.
ChatResponseUpdate responseUpdate = new(ChatRole.Assistant, [])
{
CreatedAt = generateResult.CreateTime is { } dt ? new DateTimeOffset(dt) : null,
ModelId = !string.IsNullOrWhiteSpace(generateResult.ModelVersion) ? generateResult.ModelVersion : modelId,
RawRepresentation = generateResult,
ResponseId = generateResult.ResponseId,
};
// Populate the response update contents.
responseUpdate.FinishReason = PopulateResponseContents(generateResult, responseUpdate.Contents);
// Populate usage information if there is any.
if (generateResult.UsageMetadata is { } usageMetadata)
{
responseUpdate.Contents.Add(new UsageContent(ExtractUsageDetails(usageMetadata)));
}
// Yield the update.
yield return responseUpdate;
}
}
/// <inheritdoc />
public object? GetService(System.Type serviceType, object? serviceKey = null)
{
Utilities.ThrowIfNull(serviceType, nameof(serviceType));
if (serviceKey is null)
{
// If there's a request for metadata, lazily-initialize it and return it. We don't need to worry about race conditions,
// as there's no requirement that the same instance be returned each time, and creation is idempotent.
if (serviceType == typeof(ChatClientMetadata))
{
return this._metadata ??= new("gcp.gen_ai", new("https://generativelanguage.googleapis.com/"), defaultModelId: this._defaultModelId);
}
// Allow a consumer to "break glass" and access the underlying client if they need it.
if (serviceType.IsInstanceOfType(this._models))
{
return this._models;
}
if (this._client is not null && serviceType.IsInstanceOfType(this._client))
{
return this._client;
}
if (serviceType.IsInstanceOfType(this))
{
return this;
}
}
return null;
}
/// <inheritdoc />
void IDisposable.Dispose() { /* nop */ }
/// <summary>Creates the message parameters for <see cref="Models.GenerateContentAsync(string, List{Content}, GenerateContentConfig?)"/> from <paramref name="messages"/> and <paramref name="options"/>.</summary>
private (string? ModelId, List<Content> Contents, GenerateContentConfig Config) CreateRequest(IEnumerable<ChatMessage> messages, ChatOptions? options)
{
// Create the GenerateContentConfig object. If the options contains a RawRepresentationFactory, try to use it to
// create the request instance, allowing the caller to populate it with GenAI-specific options. Otherwise, create
// a new instance directly.
string? model = this._defaultModelId;
List<Content> contents = [];
GenerateContentConfig config = options?.RawRepresentationFactory?.Invoke(this) as GenerateContentConfig ?? new();
if (options is not null)
{
if (options.FrequencyPenalty is { } frequencyPenalty)
{
config.FrequencyPenalty ??= frequencyPenalty;
}
if (options.Instructions is { } instructions)
{
((config.SystemInstruction ??= new()).Parts ??= []).Add(new() { Text = instructions });
}
if (options.MaxOutputTokens is { } maxOutputTokens)
{
config.MaxOutputTokens ??= maxOutputTokens;
}
if (!string.IsNullOrWhiteSpace(options.ModelId))
{
model = options.ModelId;
}
if (options.PresencePenalty is { } presencePenalty)
{
config.PresencePenalty ??= presencePenalty;
}
if (options.Seed is { } seed)
{
config.Seed ??= (int)seed;
}
if (options.StopSequences is { } stopSequences)
{
(config.StopSequences ??= []).AddRange(stopSequences);
}
if (options.Temperature is { } temperature)
{
config.Temperature ??= temperature;
}
if (options.TopP is { } topP)
{
config.TopP ??= topP;
}
if (options.TopK is { } topK)
{
config.TopK ??= topK;
}
// Populate tools. Each kind of tool is added on its own, except for function declarations,
// which are grouped into a single FunctionDeclaration.
List<FunctionDeclaration>? functionDeclarations = null;
if (options.Tools is { } tools)
{
foreach (var tool in tools)
{
switch (tool)
{
case AIFunctionDeclaration af:
functionDeclarations ??= [];
functionDeclarations.Add(new()
{
Name = af.Name,
Description = af.Description ?? "",
ParametersJsonSchema = af.JsonSchema,
});
break;
case HostedCodeInterpreterTool:
(config.Tools ??= []).Add(new() { CodeExecution = new() });
break;
case HostedFileSearchTool:
(config.Tools ??= []).Add(new() { Retrieval = new() });
break;
case HostedWebSearchTool:
(config.Tools ??= []).Add(new() { GoogleSearch = new() });
break;
}
}
}
if (functionDeclarations is { Count: > 0 })
{
Tool functionTools = new();
(functionTools.FunctionDeclarations ??= []).AddRange(functionDeclarations);
(config.Tools ??= []).Add(functionTools);
}
// Transfer over the tool mode if there are any tools.
if (options.ToolMode is { } toolMode && config.Tools?.Count > 0)
{
switch (toolMode)
{
case NoneChatToolMode:
config.ToolConfig = new() { FunctionCallingConfig = new() { Mode = FunctionCallingConfigMode.NONE } };
break;
case AutoChatToolMode:
config.ToolConfig = new() { FunctionCallingConfig = new() { Mode = FunctionCallingConfigMode.AUTO } };
break;
case RequiredChatToolMode required:
config.ToolConfig = new() { FunctionCallingConfig = new() { Mode = FunctionCallingConfigMode.ANY } };
if (required.RequiredFunctionName is not null)
{
((config.ToolConfig.FunctionCallingConfig ??= new()).AllowedFunctionNames ??= []).Add(required.RequiredFunctionName);
}
break;
}
}
// Set the response format if specified.
if (options.ResponseFormat is ChatResponseFormatJson responseFormat)
{
config.ResponseMimeType = "application/json";
if (responseFormat.Schema is { } schema)
{
config.ResponseJsonSchema = schema;
}
}
}
// Transfer messages to request, handling system messages specially
Dictionary<string, string>? callIdToFunctionNames = null;
foreach (var message in messages)
{
if (message.Role == ChatRole.System)
{
string instruction = message.Text;
if (!string.IsNullOrWhiteSpace(instruction))
{
((config.SystemInstruction ??= new()).Parts ??= []).Add(new() { Text = instruction });
}
continue;
}
Content content = new() { Role = message.Role == ChatRole.Assistant ? "model" : "user" };
content.Parts ??= [];
AddPartsForAIContents(ref callIdToFunctionNames, message.Contents, content.Parts);
contents.Add(content);
}
// Make sure the request contains at least one content part (the request would always fail if empty).
if (!contents.SelectMany(c => c.Parts ?? Enumerable.Empty<Part>()).Any())
{
contents.Add(new() { Role = "user", Parts = new() { { new() { Text = "" } } } });
}
return (model, contents, config);
}
/// <summary>Creates <see cref="Part"/>s for <paramref name="contents"/> and adds them to <paramref name="parts"/>.</summary>
private static void AddPartsForAIContents(ref Dictionary<string, string>? callIdToFunctionNames, IList<AIContent> contents, List<Part> parts)
{
for (int i = 0; i < contents.Count; i++)
{
var content = contents[i];
byte[]? thoughtSignature = null;
if (content is not TextReasoningContent { ProtectedData: not null } &&
i + 1 < contents.Count &&
contents[i + 1] is TextReasoningContent nextReasoning &&
string.IsNullOrWhiteSpace(nextReasoning.Text) &&
nextReasoning.ProtectedData is { } protectedData)
{
i++;
thoughtSignature = Convert.FromBase64String(protectedData);
}
Part? part = null;
switch (content)
{
case TextContent textContent:
part = new() { Text = textContent.Text };
break;
case TextReasoningContent reasoningContent:
part = new()
{
Thought = true,
Text = !string.IsNullOrWhiteSpace(reasoningContent.Text) ? reasoningContent.Text : null,
ThoughtSignature = reasoningContent.ProtectedData is not null ? Convert.FromBase64String(reasoningContent.ProtectedData) : null,
};
break;
case DataContent dataContent:
part = new()
{
InlineData = new()
{
MimeType = dataContent.MediaType,
Data = dataContent.Data.ToArray(),
DisplayName = dataContent.Name,
}
};
break;
case UriContent uriContent:
part = new()
{
FileData = new()
{
FileUri = uriContent.Uri.AbsoluteUri,
MimeType = uriContent.MediaType,
}
};
break;
case FunctionCallContent functionCallContent:
(callIdToFunctionNames ??= [])[functionCallContent.CallId] = functionCallContent.Name;
callIdToFunctionNames[""] = functionCallContent.Name; // track last function name in case calls don't have IDs
part = new()
{
FunctionCall = new()
{
Id = functionCallContent.CallId,
Name = functionCallContent.Name,
Args = functionCallContent.Arguments is null ? null : functionCallContent.Arguments as Dictionary<string, object> ?? new(functionCallContent.Arguments!),
}
};
break;
case FunctionResultContent functionResultContent:
part = new()
{
FunctionResponse = new()
{
Id = functionResultContent.CallId,
Name = callIdToFunctionNames?.TryGetValue(functionResultContent.CallId, out string? functionName) is true || callIdToFunctionNames?.TryGetValue("", out functionName) is true ?
functionName :
null,
Response = functionResultContent.Result is null ? null : new() { ["result"] = functionResultContent.Result },
}
};
break;
}
if (part is not null)
{
part.ThoughtSignature ??= thoughtSignature;
parts.Add(part);
}
}
}
/// <summary>Creates <see cref="AIContent"/>s for <paramref name="parts"/> and adds them to <paramref name="contents"/>.</summary>
private static void AddAIContentsForParts(List<Part> parts, IList<AIContent> contents)
{
foreach (var part in parts)
{
AIContent? content = null;
if (!string.IsNullOrEmpty(part.Text))
{
content = part.Thought is true ?
new TextReasoningContent(part.Text) :
new TextContent(part.Text);
}
else if (part.InlineData is { } inlineData)
{
content = new DataContent(inlineData.Data, inlineData.MimeType ?? "application/octet-stream")
{
Name = inlineData.DisplayName,
};
}
else if (part.FileData is { FileUri: not null } fileData)
{
content = new UriContent(new Uri(fileData.FileUri), fileData.MimeType ?? "application/octet-stream");
}
else if (part.FunctionCall is { Name: not null } functionCall)
{
content = new FunctionCallContent(functionCall.Id ?? "", functionCall.Name, functionCall.Args!);
}
else if (part.FunctionResponse is { } functionResponse)
{
content = new FunctionResultContent(
functionResponse.Id ?? "",
functionResponse.Response?.TryGetValue("output", out var output) is true ? output :
functionResponse.Response?.TryGetValue("error", out var error) is true ? error :
null);
}
if (content is not null)
{
content.RawRepresentation = part;
contents.Add(content);
if (part.ThoughtSignature is { } thoughtSignature)
{
contents.Add(new TextReasoningContent(null)
{
ProtectedData = Convert.ToBase64String(thoughtSignature),
});
}
}
}
}
private static ChatFinishReason? PopulateResponseContents(GenerateContentResponse generateResult, IList<AIContent> responseContents)
{
ChatFinishReason? finishReason = null;
// Populate the response messages. There should only be at most one candidate, but if there are more, ignore all but the first.
if (generateResult.Candidates is { Count: > 0 } &&
generateResult.Candidates[0] is { Content: { } candidateContent } candidate)
{
// Grab the finish reason if one exists.
finishReason = ConvertFinishReason(candidate.FinishReason);
// Add all of the response content parts as AIContents.
if (candidateContent.Parts is { } parts)
{
AddAIContentsForParts(parts, responseContents);
}
// Add any citation metadata.
if (candidate.CitationMetadata is { Citations: { Count: > 0 } citations } &&
responseContents.OfType<TextContent>().FirstOrDefault() is TextContent textContent)
{
foreach (var citation in citations)
{
textContent.Annotations =
[
new CitationAnnotation()
{
Title = citation.Title,
Url = Uri.TryCreate(citation.Uri, UriKind.Absolute, out Uri? uri) ? uri : null,
AnnotatedRegions =
[
new TextSpanAnnotatedRegion()
{
StartIndex = citation.StartIndex,
EndIndex = citation.EndIndex,
}
],
}
];
}
}
}
// Populate error information if there is any.
if (generateResult.PromptFeedback is { } promptFeedback)
{
responseContents.Add(new ErrorContent(promptFeedback.BlockReasonMessage));
}
return finishReason;
}
/// <summary>Creates an M.E.AI <see cref="ChatFinishReason"/> from a Google <see cref="FinishReason"/>.</summary>
private static ChatFinishReason? ConvertFinishReason(FinishReason? finishReason)
{
return finishReason switch
{
null => null,
FinishReason.MAX_TOKENS =>
ChatFinishReason.Length,
FinishReason.MALFORMED_FUNCTION_CALL or
FinishReason.UNEXPECTED_TOOL_CALL =>
ChatFinishReason.ToolCalls,
FinishReason.FINISH_REASON_UNSPECIFIED or
FinishReason.STOP =>
ChatFinishReason.Stop,
_ => ChatFinishReason.ContentFilter,
};
}
/// <summary>Creates a <see cref="UsageDetails"/> populated from the supplied <paramref name="usageMetadata"/>.</summary>
private static UsageDetails ExtractUsageDetails(GenerateContentResponseUsageMetadata usageMetadata)
{
UsageDetails details = new()
{
InputTokenCount = usageMetadata.PromptTokenCount,
OutputTokenCount = usageMetadata.CandidatesTokenCount,
TotalTokenCount = usageMetadata.TotalTokenCount,
};
AddIfPresent(nameof(usageMetadata.CachedContentTokenCount), usageMetadata.CachedContentTokenCount);
AddIfPresent(nameof(usageMetadata.ThoughtsTokenCount), usageMetadata.ThoughtsTokenCount);
AddIfPresent(nameof(usageMetadata.ToolUsePromptTokenCount), usageMetadata.ToolUsePromptTokenCount);
return details;
void AddIfPresent(string key, int? value)
{
if (value is int i)
{
(details.AdditionalCounts ??= [])[key] = i;
}
}
}
}
@@ -1,36 +0,0 @@
// Copyright (c) Microsoft. All rights reserved.
using Google.Apis.Util;
using Google.GenAI;
namespace Microsoft.Extensions.AI;
/// <summary>Provides implementations of Microsoft.Extensions.AI abstractions based on <see cref="Client"/>.</summary>
public static class GoogleGenAIExtensions
{
/// <summary>
/// Creates an <see cref="IChatClient"/> wrapper around the specified <see cref="Client"/>.
/// </summary>
/// <param name="client">The <see cref="Client"/> to wrap.</param>
/// <param name="defaultModelId">The default model ID to use for chat requests if not specified in <see cref="ChatOptions.ModelId"/>.</param>
/// <returns>An <see cref="IChatClient"/> that wraps the specified client.</returns>
/// <exception cref="ArgumentNullException"><paramref name="client"/> is <see langword="null"/>.</exception>
public static IChatClient AsIChatClient(this Client client, string? defaultModelId = null)
{
Utilities.ThrowIfNull(client, nameof(client));
return new GoogleGenAIChatClient(client, defaultModelId);
}
/// <summary>
/// Creates an <see cref="IChatClient"/> wrapper around the specified <see cref="Models"/>.
/// </summary>
/// <param name="models">The <see cref="Models"/> client to wrap.</param>
/// <param name="defaultModelId">The default model ID to use for chat requests if not specified in <see cref="ChatOptions.ModelId"/>.</param>
/// <returns>An <see cref="IChatClient"/> that wraps the specified <see cref="Models"/> client.</returns>
/// <exception cref="ArgumentNullException"><paramref name="models"/> is <see langword="null"/>.</exception>
public static IChatClient AsIChatClient(this Models models, string? defaultModelId = null)
{
Utilities.ThrowIfNull(models, nameof(models));
return new GoogleGenAIChatClient(models, defaultModelId);
}
}
@@ -14,8 +14,6 @@ string apiKey = Environment.GetEnvironmentVariable("GOOGLE_GENAI_API_KEY") ?? th
string model = Environment.GetEnvironmentVariable("GOOGLE_GENAI_MODEL") ?? "gemini-2.5-flash";
// Using a Google GenAI IChatClient implementation
// Until the PR https://github.com/googleapis/dotnet-genai/pull/81 is not merged this option
// requires usage of also both GeminiChatClient.cs and GoogleGenAIExtensions.cs polyfills to work.
ChatClientAgent agentGenAI = new(
new Client(vertexAI: false, apiKey: apiKey).AsIChatClient(model),
@@ -25,12 +25,7 @@ $env:GOOGLE_GENAI_MODEL="gemini-2.5-fast" # Optional, defaults to gemini-2.5-fa
### Google GenAI (Official)
The official Google GenAI package provides direct access to Google's Generative AI models. This sample uses an extension method to convert the Google client to an `IChatClient`.
> [!NOTE]
> Until PR [googleapis/dotnet-genai#81](https://github.com/googleapis/dotnet-genai/pull/81) is merged, this option requires the additional `GeminiChatClient.cs` and `GoogleGenAIExtensions.cs` files included in this sample.
>
> We appreciate any community push by liking and commenting in the above PR to get it merged and release as part of official Google GenAI package.
The official Google GenAI package provides direct access to Google's Generative AI models. This sample uses the `AsIChatClient()` extension method to convert the Google client to an `IChatClient`.
### Mscc.GenerativeAI.Microsoft (Community)
@@ -12,7 +12,7 @@ var modelPath = Environment.GetEnvironmentVariable("ONNX_MODEL_PATH") ?? throw n
// Get a chat client for ONNX and use it to construct an AIAgent.
using OnnxRuntimeGenAIChatClient chatClient = new(modelPath);
AIAgent agent = chatClient.CreateAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
AIAgent agent = chatClient.AsAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
// Invoke the agent and output the text result.
Console.WriteLine(await agent.RunAsync("Tell me a joke about a pirate."));
@@ -11,7 +11,7 @@ var modelName = Environment.GetEnvironmentVariable("OLLAMA_MODEL_NAME") ?? throw
// Get a chat client for Ollama and use it to construct an AIAgent.
AIAgent agent = new OllamaApiClient(new Uri(endpoint), modelName)
.CreateAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
.AsAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
// Invoke the agent and output the text result.
Console.WriteLine(await agent.RunAsync("Tell me a joke about a pirate."));
@@ -13,7 +13,7 @@ var model = Environment.GetEnvironmentVariable("OPENAI_MODEL") ?? "gpt-4o-mini";
AIAgent agent = new OpenAIClient(
apiKey)
.GetChatClient(model)
.CreateAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
.AsAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
// Invoke the agent and output the text result.
Console.WriteLine(await agent.RunAsync("Tell me a joke about a pirate."));
@@ -12,7 +12,7 @@ var model = Environment.GetEnvironmentVariable("OPENAI_MODEL") ?? "gpt-4o-mini";
AIAgent agent = new OpenAIClient(
apiKey)
.GetResponsesClient(model)
.CreateAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
.AsAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
// Invoke the agent and output the text result.
Console.WriteLine(await agent.RunAsync("Tell me a joke about a pirate."));
@@ -11,7 +11,7 @@ var apiKey = Environment.GetEnvironmentVariable("ANTHROPIC_API_KEY") ?? throw ne
var model = Environment.GetEnvironmentVariable("ANTHROPIC_MODEL") ?? "claude-haiku-4-5";
AIAgent agent = new AnthropicClient(new ClientOptions { APIKey = apiKey })
.CreateAIAgent(model: model, instructions: "You are good at telling jokes.", name: "Joker");
.AsAIAgent(model: model, instructions: "You are good at telling jokes.", name: "Joker");
// Invoke the agent and output the text result.
var response = await agent.RunAsync("Tell me a joke about a pirate.");
@@ -14,7 +14,7 @@ var maxTokens = 4096;
var thinkingTokens = 2048;
var agent = new AnthropicClient(new ClientOptions { APIKey = apiKey })
.CreateAIAgent(
.AsAIAgent(
model: model,
clientFactory: (chatClient) => chatClient
.AsBuilder()
@@ -23,7 +23,7 @@ AITool tool = AIFunctionFactory.Create(GetWeather);
// Get anthropic client to create agents.
AIAgent agent = new AnthropicClient { APIKey = apiKey }
.CreateAIAgent(model: model, instructions: AssistantInstructions, name: AssistantName, tools: [tool]);
.AsAIAgent(model: model, instructions: AssistantInstructions, name: AssistantName, tools: [tool]);
// Non-streaming agent interaction with function tools.
AgentThread thread = await agent.GetNewThreadAsync();
@@ -30,7 +30,7 @@ AIAgent agent = new AzureOpenAIClient(
new Uri(endpoint),
new AzureCliCredential())
.GetChatClient(deploymentName)
.CreateAIAgent(new ChatClientAgentOptions
.AsAIAgent(new ChatClientAgentOptions
{
ChatOptions = new() { Instructions = "You are good at telling jokes." },
Name = "Joker",
@@ -28,7 +28,7 @@ AIAgent agent = new AzureOpenAIClient(
new Uri(endpoint),
new AzureCliCredential())
.GetChatClient(deploymentName)
.CreateAIAgent(new ChatClientAgentOptions()
.AsAIAgent(new ChatClientAgentOptions()
{
ChatOptions = new() { Instructions = "You are a friendly travel assistant. Use known memories about the user when responding, and do not invent details." },
AIContextProviderFactory = (ctx, ct) => new ValueTask<AIContextProvider>(ctx.SerializedState.ValueKind is not JsonValueKind.Null and not JsonValueKind.Undefined
@@ -30,7 +30,7 @@ ChatClient chatClient = new AzureOpenAIClient(
// and preferably shared between multiple threads used by the same user, ensure that the
// factory reads the user id from the current context and scopes the memory component
// and its storage to that user id.
AIAgent agent = chatClient.CreateAIAgent(new ChatClientAgentOptions()
AIAgent agent = chatClient.AsAIAgent(new ChatClientAgentOptions()
{
ChatOptions = new() { Instructions = "You are a friendly assistant. Always address the user by their name." },
AIContextProviderFactory = (ctx, ct) => new ValueTask<AIContextProvider>(new UserInfoMemory(chatClient.AsIChatClient(), ctx.SerializedState, ctx.JsonSerializerOptions))
@@ -12,7 +12,7 @@ var model = Environment.GetEnvironmentVariable("OPENAI_MODEL") ?? "gpt-4o-mini";
AIAgent agent = new OpenAIClient(apiKey)
.GetChatClient(model)
.CreateAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
.AsAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
UserChatMessage chatMessage = new("Tell me a joke about a pirate.");
@@ -59,7 +59,7 @@ TextSearchProviderOptions textSearchOptions = new()
// Create the AI agent with the TextSearchProvider as the AI context provider.
AIAgent agent = azureOpenAIClient
.GetChatClient(deploymentName)
.CreateAIAgent(new ChatClientAgentOptions
.AsAIAgent(new ChatClientAgentOptions
{
ChatOptions = new() { Instructions = "You are a helpful support specialist for Contoso Outdoors. Answer questions using the provided context and cite the source document when available." },
AIContextProviderFactory = (ctx, ct) => new ValueTask<AIContextProvider>(new TextSearchProvider(SearchAdapter, ctx.SerializedState, ctx.JsonSerializerOptions, textSearchOptions)),
@@ -68,7 +68,7 @@ TextSearchProviderOptions textSearchOptions = new()
// Create the AI agent with the TextSearchProvider as the AI context provider.
AIAgent agent = azureOpenAIClient
.GetChatClient(deploymentName)
.CreateAIAgent(new ChatClientAgentOptions
.AsAIAgent(new ChatClientAgentOptions
{
ChatOptions = new() { Instructions = "You are a helpful support specialist for the Microsoft Agent Framework. Answer questions using the provided context and cite the source document when available. Keep responses brief." },
AIContextProviderFactory = (ctx, ct) => new ValueTask<AIContextProvider>(new TextSearchProvider(SearchAdapter, ctx.SerializedState, ctx.JsonSerializerOptions, textSearchOptions))
@@ -26,7 +26,7 @@ AIAgent agent = new AzureOpenAIClient(
new Uri(endpoint),
new AzureCliCredential())
.GetChatClient(deploymentName)
.CreateAIAgent(new ChatClientAgentOptions
.AsAIAgent(new ChatClientAgentOptions
{
ChatOptions = new() { Instructions = "You are a helpful support specialist for Contoso Outdoors. Answer questions using the provided context and cite the source document when available." },
AIContextProviderFactory = (ctx, ct) => new ValueTask<AIContextProvider>(new TextSearchProvider(MockSearchAsync, ctx.SerializedState, ctx.JsonSerializerOptions, textSearchOptions))
@@ -14,7 +14,7 @@ AIAgent agent = new AzureOpenAIClient(
new Uri(endpoint),
new AzureCliCredential())
.GetChatClient(deploymentName)
.CreateAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
.AsAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
// Invoke the agent and output the text result.
Console.WriteLine(await agent.RunAsync("Tell me a joke about a pirate."));
@@ -14,7 +14,7 @@ AIAgent agent = new AzureOpenAIClient(
new Uri(endpoint),
new AzureCliCredential())
.GetChatClient(deploymentName)
.CreateAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
.AsAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
// Invoke the agent with a multi-turn conversation, where the context is preserved in the thread object.
AgentThread thread = await agent.GetNewThreadAsync();
@@ -22,7 +22,7 @@ AIAgent agent = new AzureOpenAIClient(
new Uri(endpoint),
new AzureCliCredential())
.GetChatClient(deploymentName)
.CreateAIAgent(instructions: "You are a helpful assistant", tools: [AIFunctionFactory.Create(GetWeather)]);
.AsAIAgent(instructions: "You are a helpful assistant", tools: [AIFunctionFactory.Create(GetWeather)]);
// Non-streaming agent interaction with function tools.
Console.WriteLine(await agent.RunAsync("What is the weather like in Amsterdam?"));
@@ -27,7 +27,7 @@ AIAgent agent = new AzureOpenAIClient(
new Uri(endpoint),
new AzureCliCredential())
.GetChatClient(deploymentName)
.CreateAIAgent(instructions: "You are a helpful assistant", tools: [new ApprovalRequiredAIFunction(AIFunctionFactory.Create(GetWeather))]);
.AsAIAgent(instructions: "You are a helpful assistant", tools: [new ApprovalRequiredAIFunction(AIFunctionFactory.Create(GetWeather))]);
// Call the agent and check if there are any user input requests to handle.
AgentThread thread = await agent.GetNewThreadAsync();
@@ -21,7 +21,7 @@ ChatClient chatClient = new AzureOpenAIClient(
.GetChatClient(deploymentName);
// Create the ChatClientAgent with the specified name and instructions.
ChatClientAgent agent = chatClient.CreateAIAgent(name: "HelpfulAssistant", instructions: "You are a helpful assistant.");
ChatClientAgent agent = chatClient.AsAIAgent(name: "HelpfulAssistant", instructions: "You are a helpful assistant.");
// Set PersonInfo as the type parameter of RunAsync method to specify the expected structured output from the agent and invoke the agent with some unstructured input.
AgentResponse<PersonInfo> response = await agent.RunAsync<PersonInfo>("Please provide information about John Smith, who is a 35-year-old software engineer.");
@@ -33,7 +33,7 @@ Console.WriteLine($"Age: {response.Result.Age}");
Console.WriteLine($"Occupation: {response.Result.Occupation}");
// Create the ChatClientAgent with the specified name, instructions, and expected structured output the agent should produce.
ChatClientAgent agentWithPersonInfo = chatClient.CreateAIAgent(new ChatClientAgentOptions()
ChatClientAgent agentWithPersonInfo = chatClient.AsAIAgent(new ChatClientAgentOptions()
{
Name = "HelpfulAssistant",
ChatOptions = new() { Instructions = "You are a helpful assistant.", ResponseFormat = Microsoft.Extensions.AI.ChatResponseFormat.ForJsonSchema<PersonInfo>() }
@@ -16,7 +16,7 @@ AIAgent agent = new AzureOpenAIClient(
new Uri(endpoint),
new AzureCliCredential())
.GetChatClient(deploymentName)
.CreateAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
.AsAIAgent(instructions: "You are good at telling jokes.", name: "Joker");
// Start a new thread for the agent conversation.
AgentThread thread = await agent.GetNewThreadAsync();
@@ -27,7 +27,7 @@ AIAgent agent = new AzureOpenAIClient(
new Uri(endpoint),
new AzureCliCredential())
.GetChatClient(deploymentName)
.CreateAIAgent(new ChatClientAgentOptions
.AsAIAgent(new ChatClientAgentOptions
{
ChatOptions = new() { Instructions = "You are good at telling jokes." },
Name = "Joker",
@@ -29,7 +29,7 @@ using var tracerProvider = tracerProviderBuilder.Build();
// Create the agent, and enable OpenTelemetry instrumentation.
AIAgent agent = new AzureOpenAIClient(new Uri(endpoint), new AzureCliCredential())
.GetChatClient(deploymentName)
.CreateAIAgent(instructions: "You are good at telling jokes.", name: "Joker")
.AsAIAgent(instructions: "You are good at telling jokes.", name: "Joker")
.AsBuilder()
.UseOpenTelemetry(sourceName: sourceName)
.Build();
@@ -13,7 +13,7 @@ var deploymentName = System.Environment.GetEnvironmentVariable("AZURE_OPENAI_DEP
var agent = new AzureOpenAIClient(new Uri(endpoint), new AzureCliCredential())
.GetChatClient(deploymentName)
.CreateAIAgent(
.AsAIAgent(
name: "VisionAgent",
instructions: "You are a helpful agent that can analyze images");
@@ -21,7 +21,7 @@ AIAgent weatherAgent = new AzureOpenAIClient(
new Uri(endpoint),
new AzureCliCredential())
.GetChatClient(deploymentName)
.CreateAIAgent(
.AsAIAgent(
instructions: "You answer questions about the weather.",
name: "WeatherAgent",
description: "An agent that answers questions about the weather.",
@@ -32,7 +32,7 @@ AIAgent agent = new AzureOpenAIClient(
new Uri(endpoint),
new AzureCliCredential())
.GetChatClient(deploymentName)
.CreateAIAgent(instructions: "You are a helpful assistant who responds in French.", tools: [weatherAgent.AsAIFunction()]);
.AsAIAgent(instructions: "You are a helpful assistant who responds in French.", tools: [weatherAgent.AsAIFunction()]);
// Invoke the agent and output the text result.
Console.WriteLine(await agent.RunAsync("What is the weather like in Amsterdam?"));

Some files were not shown because too many files have changed in this diff Show More