2582 Commits

This Branch

This Branch

All Branches

• Release v0.50.2
  Mario Zechner · 2026-01-29 03:23:41 +01:00
  v0.50.2 9bb627a524
• docs: add missing changelog entries for v0.50.1..HEAD commits
  Mario Zechner · 2026-01-29 03:22:00 +01:00
  58aa7096dd
• fix: autocompleteMaxVisible setting not persisting

  Missing markModified() call caused the setting to be lost on reload.

  Mario Zechner · 2026-01-29 03:20:28 +01:00
  2ad4825653
• feat: add autocompleteMaxVisible setting for configurable dropdown height (#972)

  Fixes #916
  
  Adds an autocompleteMaxVisible setting (range 3-20, default 5) to control how many items are visible in the editor's autocomplete dropdown.
  
  Co-authored-by: Colin Mason <masonc789@gmail.com>

  Mario Zechner · 2026-01-29 03:17:11 +01:00
  bea3ab637f
• feat: add autocompleteMaxVisible setting for configurable dropdown height
  Colin Mason · 2026-01-29 03:16:52 +01:00
  b212314f45
• docs(coding-agent): add changelog entry for #1039
  Mario Zechner · 2026-01-29 02:52:43 +01:00
  06a7fedda5
• create headerContainer for custom header (#1039)
  Ted · 2026-01-29 02:52:08 +01:00
  d88087334f
• A couple of autocomplete improvements (#1024)

  * fix(tui): keep file suggestions open when typing in Tab-triggered mode
  
  Previously, pressing Tab on an empty prompt or after a space would show
  file suggestions, but typing a letter would dismiss them because the
  text didn't look like a path pattern. Now the editor tracks whether
  autocomplete was triggered via Tab (force mode) or naturally (regular
  mode), and uses the appropriate suggestion method when updating.
  
  * fix(tui): hide autocomplete when backspacing slash command to empty
  
  Previously, typing / showed slash command suggestions, but pressing
  Backspace to delete it showed file suggestions instead of hiding all
  suggestions.

  Sviatoslav Abakumov · 2026-01-29 02:48:09 +01:00
  b54d689ec1
• fix(tui): remove backslash input buffering (#1037)

  Instead of buffering `\` and waiting for the next key, let it be
  inserted immediately. On Enter, check if preceded by `\` and treat as
  newline.
  
  Removed backslash handling from the Input component entirely.

  Sviatoslav Abakumov · 2026-01-29 02:47:15 +01:00
  d57a26c88b
• fix: preserve external settings.json edits on reload (#1046)

  Co-authored-by: Mario Zechner <badlogicgames@gmail.com>

  Nico Bailon · 2026-01-29 02:42:23 +01:00
  bac57f81be
• feat(ai): add Hugging Face provider support

  - Add huggingface to KnownProvider type
  - Add HF_TOKEN env var mapping
  - Process huggingface models from models.dev (14 models)
  - Use openai-completions API with compat settings
  - Add tests for all provider test suites
  - Update documentation
  
  fixes #994

  Mario Zechner · 2026-01-29 02:40:14 +01:00
  c808de605a
• feat: add /files command to list session file operations
  Mario Zechner · 2026-01-29 02:25:59 +01:00
  f3cfb7e1ae
• feat(ai): add PI_CACHE_RETENTION env var for extended prompt caching

  Adds support for extended cache retention via PI_CACHE_RETENTION=long:
  - Anthropic: 5m -> 1h TTL
  - OpenAI: in-memory -> 24h retention
  
  Only applies to direct API calls (api.anthropic.com, api.openai.com).
  Proxies and other providers are unaffected.
  
  fixes #967

  Mario Zechner · 2026-01-29 02:22:06 +01:00
  1b6a147579
• fix(ai): normalize pipe-separated tool call IDs for cross-provider handoff

  - Handle pipe-separated IDs from OpenAI Responses API in openai-completions provider
  - Strip trailing underscores after truncation in openai-responses-shared (OpenAI Codex rejects them)
  - Add regression tests for tool call ID normalization
  
  fixes #1022

  Mario Zechner · 2026-01-29 01:28:12 +01:00
  605f6f494b
• fix(coding-agent): reset retry counter after each successful LLM response

  Previously, within a single tool-use turn, rate limit retries would
  accumulate across separate LLM calls. For example, if each of 3 tool
  calls hit a 429 and retried once, the counter would show '3/3' and fail
  even though each individual retry succeeded.
  
  Now the counter resets immediately when a successful (non-error)
  assistant message arrives, so each LLM call gets a fresh set of retries.
  
  Fixes #1019

  Mario Zechner · 2026-01-29 01:12:29 +01:00
  4f004adefa
• fix(ai): 429 rate limit errors no longer trigger auto-compaction

  429 (Too Many Requests) was incorrectly classified as context overflow,
  triggering compaction instead of retry with backoff. The original logic
  assumed token-based rate limiting correlates with context overflow, but
  these are different concepts:
  
  - Rate limiting (429): requests/tokens per time period (throughput)
  - Context overflow: single request exceeds context window (size)
  
  Now 429 errors are handled by the existing retry logic with exponential
  backoff, while 400/413 remain as potential context overflow indicators.
  
  fixes #1038

  Mario Zechner · 2026-01-29 00:43:38 +01:00
  25707f9ad4
• fix(coding-agent): improve thinking defaults in footer
  Mario Zechner · 2026-01-29 00:33:45 +01:00
  a373dce747
• feat(tui,coding-agent): add shell-style keybindings alt+b, alt+f, ctrl+d

  - Add alt+b/alt+f as alternative bindings for cursorWordLeft/cursorWordRight
  - Add ctrl+d as alternative binding for deleteCharForward
  - Fix ctrl+d in custom editor to perform delete when text is present
    (previously it was always consumed, even with non-empty input)
  
  Closes #1043
  
  Co-authored-by: Jason Ish <ish@unx.ca>

  Mario Zechner · 2026-01-29 00:21:39 +01:00
  81be81328c
• Closes #1041, incorrect .md file name in warning messages.
  Mario Zechner · 2026-01-29 00:16:50 +01:00
  5ef79459dc
• fix(coding-agent): empty array in package filter now disables all resources

  fixes #1044

  Mario Zechner · 2026-01-29 00:12:42 +01:00
  d21cbde87f
• fix(ai): preserve input token counts from message_start in Anthropic provider

  Proxies like Portkey omit input_tokens in message_delta events (it's nullable
  per the SDK). The previous code unconditionally overwrote usage fields, causing
  input token counts to reset to 0.
  
  Now only updates usage fields when they are present (not null), preserving
  the correct input_tokens value captured from message_start.
  
  Fixes #1045

  Mario Zechner · 2026-01-29 00:06:51 +01:00
  8b5c81f21f
• fix(ai): detect DeepSeek URLs and disable unsupported developer role

  fixes #1048

  Mario Zechner · 2026-01-28 23:55:54 +01:00
  4f9deddd47
• fix(ai): handle sensitive stop_reason from Anthropic API (fixes #978)
  mom · 2026-01-28 02:18:16 +00:00
  ee7c0a7d18
• OpenAI completions toolChoice fix (#998)

  * openai completions tools fix
  
  * Reset generated file
  
  ---------
  
  Co-authored-by: williamtwomey <ai@shadylawn.net>

  williamtwomey · 2026-01-28 03:03:15 +01:00
  41d2c7ff38
• feat(web-ui): export CustomProviderCard, ProviderKeyInput, AbortedMessage, ToolMessageDebugView (fixes #1015)
  mom · 2026-01-28 02:02:14 +00:00
  dd854c4a75
• Revert "feat(web-ui): export CustomProviderCard, ProviderKeyInput, AbortedMessage, ToolMessageDebugView (fixes #1015)"

  This reverts commit ec0aa5c30e.

  mom · 2026-01-28 02:00:50 +00:00
  7347d755e4
• feat(web-ui): export CustomProviderCard, ProviderKeyInput, AbortedMessage, ToolMessageDebugView (fixes #1015)
  mom · 2026-01-28 02:00:07 +00:00
  ec0aa5c30e
• fix: typos in comments (fixes #1016)
  mom · 2026-01-28 01:56:34 +00:00
  ebd8e895ce
• fix(tui): improve table rendering with row dividers and min width (#997)

  * fix(tui): improve table rendering
  
  * fix(tui): handle narrow table widths

  Thomas Mustier · 2026-01-28 02:42:08 +01:00
  e7b9209daf
• perf(tui): scan only bottom terminal height lines in extractCursorPosition
  mom · 2026-01-28 01:40:57 +00:00
  fb693fbc90
• perf(tui): early-out extractCursorPosition when cursor is above visible viewport
  mom · 2026-01-28 01:37:43 +00:00
  9d47747ff6
• perf(tui): optimized extractCursorPosition to scan lines in reverse order (#1004)

  - Optimized extractCursorPosition to scan lines in reverse order for improved performance.

  Can Bölük · 2026-01-28 02:34:50 +01:00
  a8f33fd630
• Add get_commands RPC for headless clients (#995)

  * Add get_commands RPC for headless clients
  
  Headless clients like Emacs can now query which commands are available.
  Previously they could only discover file-based prompt templates by
  scanning the filesystem; extension commands and skills were invisible.
  
  The response includes each command's name, description, and source
  (extension, template, or skill). Commands appear in the same order
  as the TUI's autocomplete: extension commands first, then templates,
  then skills.
  
  Built-in TUI commands (/settings, /fork, etc.) are excluded since
  they require the interactive UI. Commands like /compact have dedicated
  RPC equivalents instead.
  
  * Add location and path to get_commands response
  
  Clients can show where commands come from (user/project/path) and
  display file paths in tooltips. The data is already available on
  templates and skills - just exposing it.

  Daniel Nouri · 2026-01-28 02:34:15 +01:00
  0ad189f12a
• fix(coding-agent): hide provider in footer when terminal is narrow (#981)
  Sviatoslav Abakumov · 2026-01-28 02:15:27 +01:00
  c67b582fc4
• feat(tui): auto-apply single suggestion in force file autocomplete (#993)

  When Tab triggers file autocomplete and there's exactly one matching
  suggestion, apply it immediately without showing the menu. This makes
  path completion faster by eliminating the extra Tab press to confirm.

  Sviatoslav Abakumov · 2026-01-28 02:12:18 +01:00
  1224b31135
• Add [Unreleased] section for next cycle
  Mario Zechner · 2026-01-27 00:15:42 +01:00
  2cc2544809
• Release v0.50.1
  Mario Zechner · 2026-01-27 00:14:58 +01:00
  v0.50.1 29c59d79ba
• docs(coding-agent): clarify dependency categories
  Mario Zechner · 2026-01-27 00:13:54 +01:00
  bec72c428c
• docs(coding-agent): clarify pi package dependency bundling
  Mario Zechner · 2026-01-27 00:10:58 +01:00
  484b605ca0
• chore: update changelogs
  Mario Zechner · 2026-01-27 00:08:06 +01:00
  6159eb4be8
• fix(coding-agent): prune empty git dirs on remove
  Mario Zechner · 2026-01-26 23:58:14 +01:00
  ee7d03cf52
• fix(ai): filter deprecated OpenCode models from generation (#970)

  Add status === 'deprecated' check for OpenCode Zen models, matching
  the existing pattern used for GitHub Copilot models. This removes
  deprecated models like glm-4.7-free and minimax-m2.1-free from the
  generated model catalog.

  Daniel Tatarkin · 2026-01-26 23:56:13 +01:00
  9f3eef65f8
• fix(diff): keep picker open after selection
  Mario Zechner · 2026-01-26 23:04:32 +01:00
  a5f603d704
• fix(coding-agent): bind extension UI context on startup
  Mario Zechner · 2026-01-26 23:00:48 +01:00
  5dbeadae05
• fix(coding-agent): sync agent state after newSession setup callback

  Move setup callback handling from interactive/rpc modes into AgentSession.newSession().
  After setup() runs, sync agent state via replaceMessages() so the LLM has context
  and the UI renders the messages properly.
  
  fixes #968

  Mario Zechner · 2026-01-26 22:25:09 +01:00
  f86e3c3103
• fix(coding-agent): handle force-pushed git extensions in pi update

  Simplified to use @{upstream} instead of branch detection.
  Moved changelog entry to [Unreleased].
  
  Closes #961

  Mario Zechner · 2026-01-26 22:02:17 +01:00
  0134a2a73d
• fix(coding-agent): handle force-pushed git extensions in pi update
  Aliou Diallo · 2026-01-26 22:01:08 +01:00
  bfb21e31f8
• docs(coding-agent): add placement note for extension hot-reload
  Mario Zechner · 2026-01-26 21:55:57 +01:00
  3c252d50f5
• fix(coding-agent): Title Case extension hotkeys in /hotkeys output (#969)
  Sviatoslav Abakumov · 2026-01-26 21:50:51 +01:00
  3ae0cbe384
• Add [Unreleased] section for next cycle
  Mario Zechner · 2026-01-26 16:55:24 +01:00
  2bbc1a7659