Fimeg/letta-server
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

Commit Graph

  • 1d1adb261a fix: orphaned approvals, token inflation, reasoning fields, memfs redis dep ani-patches Ani Tunturi 2026-03-26 23:24:32 -04:00
  • 08d3c26732 fix: sanitize control characters before sending to inference backends Ani Tunturi 2026-03-21 20:23:56 -04:00
  • 2b3cfb1797 feat: breathe — 220K context window, 25-result search pages Ani Tunturi 2026-03-21 17:41:56 -04:00
  • 9af8e94fc9 fix: use exclude_none instead of per-provider field stripping Ani Tunturi 2026-03-21 12:42:03 -04:00
  • 93337ce680 fix: pass images from tool returns to vision models Ani Tunturi 2026-03-21 12:41:37 -04:00
  • c79de81cea fix: accept base64 images in tool returns, not just letta-hosted Ani Tunturi 2026-03-21 12:41:20 -04:00
  • 4cb2f21c65 chore: bump 0.16.6 (#3211) cthomas 2026-03-03 19:13:07 -08:00
  • 9eb7ae39a2 chore: bump version Caren Thomas 2026-03-03 18:34:51 -08:00
  • aeeec41859 feat: new agent id query param for default convo (#9756) cthomas 2026-03-03 16:29:09 -08:00
  • a5bac26556 fix: add "max" to AnthropicModelSettings effort type (#9754) jnjpng 2026-03-03 16:26:45 -08:00
  • 54369a7c12 fix: add billing_context to SleeptimeMultiAgent V3/V4 stream methods (#9753) cthomas 2026-03-03 14:06:19 -08:00
  • 4e2bf3ecd6 feat: add gpt-5.3-chat-latest model support (#9746) Kevin Lin 2026-03-03 13:09:06 -08:00
  • 774305d10a feat(web): device-mode refactor to use proper websocket typing (#9740) Charles Packer 2026-03-03 13:07:48 -08:00
  • 416ffc7cd7 Add billing context to LLM telemetry traces (#9745) cthomas 2026-03-03 13:05:43 -08:00
  • db9e0f42af fix(core): prevent ModelSettings default max_output_tokens from overriding agent config (#9739) jnjpng 2026-03-03 11:48:19 -08:00
  • 8335aa0fa0 fix: add some more logging for interrupts (#9733) Ari Webb 2026-03-02 16:50:49 -08:00
  • c8ae02a1fb feat(core): sort agents by updated_at [LET-7771] (#9730) Christina Tong 2026-03-02 13:10:43 -07:00
  • c247496027 Clean up server (#9728) Shubham Naik 2026-03-02 09:19:00 -10:00
  • 8e60b73eee fix: minor change in upsert logic for prompt default (#9729) amysguan 2026-03-02 11:11:48 -08:00
  • c28ba77354 Fix: ADE compaction button compacts current conversation (#9720) amysguan 2026-03-02 10:03:02 -08:00
  • 7a4188dbda Add compaction settings to ADE (#9667) amysguan 2026-03-02 10:02:57 -08:00
  • 612a2ae98b Fix: Change Z.ai context window to account for max_token subtraction (#9710) amysguan 2026-02-27 17:10:46 -08:00
  • aa66e81a71 feat: add debug logs in telem endpoint (#9723) cthomas 2026-02-27 16:48:54 -08:00
  • a50482e6d3 feat(core): sync skills from SKILL.md into memFS blocks (#9718) Sarah Wooders 2026-02-27 14:47:14 -08:00
  • a11ba9710c feat(core): increase Gemini timeout to 10 minutes (#9714) Kevin Lin 2026-02-27 11:15:37 -08:00
  • ab784e702d feat: support default convo in list and cancel endpoints (#9707) cthomas 2026-02-26 17:17:15 -08:00
  • 39a537a9a5 feat: add default convo support to conversations endpoint (#9706) cthomas 2026-02-26 16:19:39 -08:00
  • fd4a8e73a5 More stream fixes (#9702) Shubham Naik 2026-02-26 15:40:38 -08:00
  • 673c1220a1 fix: strip properties for fireworks (#9703) Ari Webb 2026-02-26 15:12:21 -08:00
  • 57e7e0e52b feat(core): reserve skills in memfs sync and list top-level skill directory [LET-7710] (#9691) Sarah Wooders 2026-02-26 13:41:05 -08:00
  • 750b83a2ea fix: update conversation manager tests for auto-generated system message (#9687) jnjpng 2026-02-26 11:42:18 -08:00
  • 28a66fa9d7 chore: remove stmt timeout debug logging (#9693) cthomas 2026-02-26 11:26:17 -08:00
  • f54ae7c929 feat: render description for non-system files in memory_filesystem tree (#9688) github-actions[bot] 2026-02-25 18:54:58 -08:00
  • bf80de214d feat: change default context window from 32000 to 128000 (#9673) github-actions[bot] 2026-02-25 18:33:11 -08:00
  • 357a3ad15b Shub/let 7721 make env permanent [LET-7721] (#9683) Shubham Naik 2026-02-25 17:58:12 -08:00
  • 46971414a4 fix: preserve agent max_tokens when caller doesn't explicitly set it (#9679) jnjpng 2026-02-25 17:18:08 -08:00
  • 5d55d4ccd4 chore: rebuild docs (#9674) Shubham Naik 2026-02-25 13:10:06 -08:00
  • 1fb355a39a fix: override stop reason for streaming for empty response (#9663) cthomas 2026-02-25 11:00:14 -08:00
  • bd6f2e792c feat: accept recent_chunks in error telemetry schema (#9662) jnjpng 2026-02-25 10:56:47 -08:00
  • dd0e513951 fix: lazy load conversations [LET-7682] (#9629) Ari Webb 2026-02-25 10:38:13 -08:00
  • 1555c338b6 fix(core-tests): update SDK blocks limit expectation to 100k (#9653) Charles Packer 2026-02-25 00:05:40 -08:00
  • 9422b2d993 fix: set otid for all approval request message (#9655) cthomas 2026-02-24 20:23:55 -08:00
  • 1448609ecf fix: set otid for summary message (#9654) cthomas 2026-02-24 19:58:12 -08:00
  • 3d781efd21 fix(core): raise LLMEmptyResponseError for empty Anthropic responses (#9624) cthomas 2026-02-24 19:57:23 -08:00
  • 86ff216dc9 fix: update tests for CORE_MEMORY_BLOCK_CHAR_LIMIT increase to 100k (#9645) github-actions[bot] 2026-02-24 16:39:25 -08:00
  • 895acb9f4e feat(core): add gpt-5.3-codex model support (#9628) Kevin Lin 2026-02-24 14:56:56 -08:00
  • ddfa922cde fix(core): prevent event loop saturation from ClickHouse and socket trace writes (#9617) Kian Jones 2026-02-24 12:53:20 -08:00
  • 94fc05b6e5 feat: remove limit from git-base memory frontmatter and increase default to 100k (#9537) github-actions[bot] 2026-02-24 12:00:54 -08:00
  • 0020f4b866 feat: recompile system message on new conversation creation (#9508) github-actions[bot] 2026-02-24 11:48:05 -08:00
  • 1b2aa98b3e chore: bump version 0.16.5 (#3202) cthomas 2026-02-24 11:02:17 -08:00
  • ce54fb1a00 bump version Caren Thomas 2026-02-24 10:58:16 -08:00
  • 47b0c87ebe Add modes self and self_sliding_window for prompt caching (#9372) amysguan 2026-02-24 10:15:36 -08:00
  • 47d55362a4 fix: models need to be paginated (#9621) Ari Webb 2026-02-23 18:08:47 -08:00
  • 8ab9d78a23 chore: cleanup (#9602) cthomas 2026-02-23 16:49:42 -08:00
  • db418d99f4 test: remove sonnet 3-7 reference (#9618) cthomas 2026-02-23 16:42:27 -08:00
  • c325b2b002 feat: add memfs file list and read endpoints to cloud-api [LET-7437] (#9520) Ari Webb 2026-02-23 16:03:27 -08:00
  • 5505e9cf4b fix(core): suppress missing-otid warning for compaction events (#9616) jnjpng 2026-02-23 13:02:10 -08:00
  • d77c2950dc chore: update APIs for conversations (#9614) Sarah Wooders 2026-02-23 11:54:23 -08:00
  • 62967bcca0 feat: parallel tool calling minimax provider [LET-7647] (#9613) Ari Webb 2026-02-23 11:33:14 -08:00
  • a59f24ac87 fix(core): ensure buffered Anthropic tool chunks always include otid (#9516) jnjpng 2026-02-23 11:21:49 -08:00
  • f082fd5061 feat: add order_by and order params to /v1/conversations list endpoin… (#9599) Shubham Naik 2026-02-23 10:42:33 -08:00
  • afbc416972 feat(core): add model/model_settings override fields to conversation create/update (#9607) Sarah Wooders 2026-02-22 23:12:05 -08:00
  • a9a6a5f29d fix: add correct logging (#9603) Ari Webb 2026-02-20 16:36:04 -08:00
  • 8fc77af685 fix(memory): standardize tool parameter names (#9552) Kevin Lin 2026-02-20 16:14:48 -08:00
  • 9155b4fa86 fix: use canonical stop reason mapping in redis stream finalizer (#9600) jnjpng 2026-02-20 15:58:51 -08:00
  • ae24bd1034 feat: add recent_chunks to feedback endpoint for diagnostics (#9594) jnjpng 2026-02-20 15:54:58 -08:00
  • ba67621e1b feat: add conversation deletion endpoint (soft delete) [LET-7286] (#9230) github-actions[bot] 2026-02-20 15:28:25 -08:00
  • 9c8589a687 fix: correct ChatGPT OAuth GPT-5 max output token defaults (#9592) jnjpng 2026-02-20 14:52:48 -08:00
  • 73c824f5d2 feat: make agent_id optional in conversations list endpoint [LET-7612] (#9585) Shubham Naik 2026-02-20 12:29:21 -08:00
  • 257b99923b fix: preserve max_tokens on model_settings updates without max_output_tokens (#9591) jnjpng 2026-02-20 12:27:08 -08:00
  • 857c289ed2 fix: handle compact edge case in idempotency check (#9588) cthomas 2026-02-20 12:18:41 -08:00
  • 34bab3cf9a Shub/let listener mode control (#9584) Shubham Naik 2026-02-20 12:06:18 -08:00
  • 73c9b14fa9 fix: dont throw error if compaction races (#9576) cthomas 2026-02-19 21:52:51 -08:00
  • f10440b49c fix: update Anthropic Haiku test model after 3.5 retirement (#9569) jnjpng 2026-02-19 18:22:26 -08:00
  • ddaf4053f6 test: fix parallel tool call default value (#9572) cthomas 2026-02-19 18:10:15 -08:00
  • a101d5980d Fix: load config for summarizer model from defaults instead of agent's config (#9568) amysguan 2026-02-19 17:52:38 -08:00
  • b29d063ba7 feat: forward letta-code feedback context fields to PostHog (#9567) jnjpng 2026-02-19 17:16:34 -08:00
  • 33969d7190 Default to lightweight compaction model instead of agent's model (#9488) amysguan 2026-02-19 15:31:13 -08:00
  • eb4a0daabd fix: allow explicit null for max_tokens on GPT-5 models (#9562) jnjpng 2026-02-19 15:14:10 -08:00
  • 828c89c76f fix: populate max_tokens when listing LLM models (#9559) jnjpng 2026-02-19 14:02:49 -08:00
  • bd5b5fa9f3 feat(gemini): add 3.1 pro preview support (#9553) Kevin Lin 2026-02-19 12:09:59 -08:00
  • e2ad8762fe fix: gemini streaming bug (#9555) cthomas 2026-02-19 11:28:53 -08:00
  • 8ffc515674 fix: flip parallel_tool_calls setting default (#9541) cthomas 2026-02-19 10:49:35 -08:00
  • 3cdd64dc24 chore: update keepalive interval 50->20 (#9538) cthomas 2026-02-19 10:19:45 -08:00
  • 8f56527958 perf(memfs): delta upload — only push new git objects after commit (#9548) Kian Jones 2026-02-18 21:16:03 -08:00
  • 044241daec fix(core): include effort in AnthropicModelSettings returned by _to_model_settings() (#9543) Charles Packer 2026-02-18 20:13:19 -08:00
  • e65795b5f1 fix(core): handle None message_ids in context window calculator (#9330) Kian Jones 2026-02-18 18:40:26 -08:00
  • e8d5922ff9 fix(core): handle ResponseIncompleteEvent in OpenAI Responses API streaming (#9535) jnjpng 2026-02-18 15:44:22 -08:00
  • 5896e5d023 fix: logging for credit verification step (#9514) Ari Webb 2026-02-17 17:20:53 -08:00
  • 3651658ea7 fix: tool call streaming using deprecated field (#9517) cthomas 2026-02-17 16:41:37 -08:00
  • 21765d16c9 fix(core): add OpenAI 24h prompt cache retention for supported models (#9509) Ari Webb 2026-02-17 15:07:21 -08:00
  • 042c9c36af fix(core): add warning log for streaming chunks missing id or otid (#9513) jnjpng 2026-02-17 14:53:26 -08:00
  • 3247fa7065 chore: fix favoriting bugs (#9505) Shubham Naik 2026-02-17 13:24:30 -08:00
  • 4a829123cd Listener mode (#9486) Shubham Naik 2026-02-17 11:10:08 -08:00
  • f5c4ab50f4 chore: add ty + pre-commit hook and repeal even more ruff rules (#9504) Kian Jones 2026-02-17 10:49:22 -08:00
  • 39ddda81cc feat: add Anthropic Sonnet 4.6 (#9408) Devansh Jain 2026-02-17 10:13:05 -08:00
  • 25d54dd896 chore: enable F821, F401, W293 (#9503) Kian Jones 2026-02-17 10:07:40 -08:00
  • fa70e09963 Revert "fix(core): add OpenAI prompt cache key and model-gated 24h retention" (#9502) Ari Webb 2026-02-16 21:24:04 -08:00
  • 2bf3314cef fix: import asyncio for parallel tool calls (#9501) Sarah Wooders 2026-02-16 17:10:53 -08:00
  • 20c71523f8 chore: hotwire fix for core (#9500) Shubham Naik 2026-02-16 16:03:47 -08:00