Commit Graph

62 Commits

Author SHA1 Message Date
Charles Packer
a0f604b5f0 fix: prevent duplicate rendering (#528)
Co-authored-by: Letta <noreply@letta.com>
2026-01-12 20:46:57 -08:00
Charles Packer
77c51eb23d fix: retry on network errors during LLM streaming (#519)
Co-authored-by: Letta <noreply@letta.com>
2026-01-11 19:40:21 -08:00
Charles Packer
ed9f227828 fix: Fix subagents not inheriting parent's tool permission rules (#492)
Co-authored-by: Letta <noreply@letta.com>
2026-01-07 22:37:00 -08:00
Charles Packer
34367de5d7 feat: implement client-side tools via client_tools spec (#456)
Co-authored-by: Letta <noreply@letta.com>
2026-01-02 23:35:40 -08:00
Charles Packer
6587237fa6 fix: patch desync fallback (#452) 2026-01-02 16:20:40 -08:00
Charles Packer
a956a15db4 feat: add system prompt and memory block configuration for headless mode (#450)
Co-authored-by: Letta <noreply@letta.com>
2026-01-02 15:32:07 -08:00
Charles Packer
321519a1b7 fix: add retry on approval error from desync (#449) 2026-01-02 14:52:41 -08:00
Charles Packer
2c7e3aabe3 fix: patch headless permission mode (#447)
Co-authored-by: Letta <noreply@letta.com>
2026-01-02 09:50:05 -08:00
Charles Packer
397560ef00 feat: add typed wire format for stream-json protocol (#445)
Co-authored-by: Letta <noreply@letta.com>
2026-01-01 20:05:41 -08:00
Charles Packer
47e4734ba1 feat: Add --input-format stream-json for bidirectional SDK communication (#444)
Co-authored-by: Letta <noreply@letta.com>
2026-01-01 01:00:45 -08:00
Charles Packer
725a2a7cad fix: update stream-json output format for SDK compatibility (#442)
Co-authored-by: Letta <noreply@letta.com>
2025-12-31 19:19:35 -08:00
Charles Packer
aeadf27938 fix: add nested fallback retry (#432) 2025-12-30 18:27:23 -08:00
Charles Packer
f4c71adab6 feat: add automatic retry for transient LLM API errors (#389)
Co-authored-by: Letta <noreply@letta.com>
2025-12-25 09:49:19 -08:00
Devansh Jain
7d73219c50 chore: Add --system flag for existing agents + update system prompt variable names (#328) 2025-12-19 17:16:21 -08:00
Charles Packer
f38bd1b133 fix: auto-retry when tools missing on server + graceful error handling (#315)
Co-authored-by: Letta <noreply@letta.com>
2025-12-18 20:52:07 -08:00
Kian Jones
322ab19f11 feat: assorted small changes (#244)
Co-authored-by: Letta <noreply@letta.com>
2025-12-16 17:20:31 -05:00
Devansh Jain
ae54666a98 feat: Stateless subagents (#127) 2025-12-15 20:16:25 -08:00
Kevin Lin
f354e6aa41 fix: pass model and system prompt when resuming agent (#227) 2025-12-15 19:38:13 -08:00
Charles Packer
2f21893ef5 feat: profile-based persistence with startup selector (#212)
Co-authored-by: Letta <noreply@letta.com>
2025-12-15 11:13:43 -08:00
Ari Webb
82518f135b feat: add from af cli flag (#195) 2025-12-14 14:27:31 -08:00
Charles Packer
2936b57ffc feat: remove uimode/parallel from settings (#206) 2025-12-14 14:25:25 -08:00
Kevin Lin
a58dddaf3d feat: Add retries for stop_reason error with metadata (#191) 2025-12-12 09:51:49 -08:00
Charles Packer
c39bda9a01 fix: improve error handling for headless (#187) 2025-12-11 22:30:26 -08:00
Kevin Lin
d5b02158d5 feat: Exponential backoff for llm_api_error (#181) 2025-12-11 15:54:34 -08:00
Christina Tong
b6c387b9ec feat: message queueing [LET-4669] (#171) 2025-12-10 11:08:55 -08:00
Devansh Jain
38b532e8ba fix: Make --tools work (#167) 2025-12-08 18:34:00 -08:00
Kevin Lin
cf5c738eb9 fix: Handle response with only tool call or reasoning (#160) 2025-12-05 22:19:31 -08:00
Kevin Lin
2236cdbc14 fix: Enable --system flag for headless mode (#153) 2025-12-04 11:04:55 -08:00
Charles Packer
09d120a578 fix: better logging on error (#152) 2025-12-03 22:15:44 -08:00
Charles Packer
5df327755f feat: add startup status messages showing agent provenance (#147)
Co-authored-by: Letta <noreply@letta.com>
2025-12-02 00:10:45 -08:00
Kevin Lin
57169c63e1 feat: Add skill creation command (#141)
Co-authored-by: cpacker <packercharles@gmail.com>
Co-authored-by: Sarah Wooders <sarahwooders@gmail.com>
2025-12-01 21:02:49 -08:00
Charles Packer
06ec8b9d3c feat: new flag to filter blocks on creation (#132) 2025-11-27 02:12:41 -08:00
Devansh Jain
8b3523c1a3 feat: Skills omni-tool (#102) 2025-11-26 11:16:57 -08:00
Charles Packer
d002c75f9c feat: add toolset flag (#116) 2025-11-23 20:13:02 -08:00
Charles Packer
9212dfb52b feat: add --sleeptime CLI flag (#94)
Co-authored-by: Letta <noreply@letta.com>
2025-11-16 23:44:06 -08:00
Kevin Lin
754db3b273 fix: Patch headless mode GPT-5 (#88)
Co-authored-by: cpacker <packercharles@gmail.com>
2025-11-16 20:48:50 -08:00
Charles Packer
c234ea2b54 fix: parallel tool calling misc fixes (#85) 2025-11-08 12:36:52 -08:00
Charles Packer
d762859963 refactor: extract shared approval batch execution logic (#82)
Co-authored-by: Letta <noreply@letta.com>
2025-11-07 18:23:16 -08:00
Charles Packer
ac0b929119 fix: support parallel tool calling in headless mode (#81)
Co-authored-by: Letta <noreply@letta.com>
2025-11-07 18:05:27 -08:00
Charles Packer
2845781811 feat: add settings for parallelToolCalls and enableSleeptime (#80)
Co-authored-by: Letta <noreply@letta.com>
2025-11-07 17:47:58 -08:00
cpacker
182e330818 fix: patch plan mode 2025-11-07 16:17:25 -08:00
Devansh Jain
ea313159ce feat: Support for skills (#76) 2025-11-07 15:00:01 -08:00
Charles Packer
42eb671bf4 fix: if no stop reason, attempt to resume from background mode (#56) 2025-11-04 11:20:58 -08:00
Devansh Jain
5c3b303eac chore: Create new memory blocks with --new (#58) 2025-11-03 19:20:03 -08:00
Charles Packer
c5b9f50767 fix: handle possible case of desync on pending approvals (#53) 2025-11-02 22:12:12 -08:00
Charles Packer
f7cea4a830 fix: handle mid-stream errors properly using sdk typing (#54) 2025-11-02 20:36:05 -08:00
Charles Packer
3dc6963dfb fix: improve error handling in interactive and headless modes (#48)
Co-authored-by: Letta <noreply@letta.com>
2025-10-31 18:25:30 -07:00
Shubham Naik
e384672a5a Sign in with letta cloud (#44)
Co-authored-by: Shubham Naik <shub@memgpt.ai>
2025-10-31 14:47:21 -07:00
Shubham Naik
5108d1eda3 Refactor settings manager (#41)
Co-authored-by: Shubham Naik <shub@memgpt.ai>
2025-10-31 11:54:39 -07:00
Devansh Jain
32a3f7c7ab fix: Model flag parsing and model args (#42) 2025-10-30 20:23:21 -07:00