ceos/opencode
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
14,989 commits 1,157 branches 1,083 tags 3 GiB
Commit graph

62 commits

Author SHA1 Message Date
Aiden Cline
49e86dd8ef feat(llm): enforce terminal stream contract for every route 
LLMClient.stream now guarantees, for native protocol routes and the
synthetic AI SDK route alike: a successful stream emits exactly one
terminal finish event and nothing after it. EOF before finish means the
provider stream was truncated (proxy cut, silent drop) and fails as
LLM.MalformedResponse instead of letting a partial response settle as
complete; output after finish also fails. Enforcement applies after
protocol parsing so stream.onHalt flushes remain subject to it, with
per-subscription state.
 2026-07-13 12:23:39 -05:00
Aiden Cline
5d87c7ad1f refactor(llm): remove the provider-error stream event 
Delete the provider-error LLMEvent so streams carry output only and
every provider-reported failure exits through the typed error channel.
Anthropic SSE error events, OpenAI Responses response.failed/error
events, and Bedrock exception frames now fail the stream with an
LLMError from the shared classifier (carrying the provider code, e.g.
rate_limit_exceeded -> LLM.RateLimit).

Core follows: the publisher drops its provider-error case, the runner
drops held-back overflow events (overflow recovery keys off a thrown
LLM.ContextOverflow), compaction and title stop scanning events for
failures, and isContextOverflowFailure is deleted with its last
consumer. V1 packages/opencode gets minimal compile fixes only (dead
switch case, one test stream).
 2026-07-13 12:22:30 -05:00
Aiden Cline
fce506b3f9 refactor(llm): replace LLMError reasons with flat tagged union 
Replace the LLMError { module, method, reason } wrapper with a flat
tagged union (LLM.BadRequest, LLM.Authentication, LLM.PermissionDenied,
LLM.NotFound, LLM.RateLimit, LLM.QuotaExceeded, LLM.ContentPolicy,
LLM.ContextOverflow, LLM.ServerError, LLM.APIError, LLM.ConnectionError,
LLM.TimeoutError, LLM.MalformedResponse, LLM.NoRoute) plus an isLLMError
guard. Add one shared classifyApiFailure classifier used by the HTTP
executor and the AI SDK adapter so both surfaces classify identically,
preserving status, headers, body, and retry-after.

Core policy moves onto tags: retry RateLimit | ServerError |
ConnectionError | TimeoutError; toSessionError adds
provider.context-overflow, provider.timeout, and provider.not-found.

The provider-error stream event and the runner's held-back overflow
handling are unchanged here; isContextOverflowFailure now bridges old
events and new tags until the event is removed.
 2026-07-13 12:21:21 -05:00
Harsh Mathur
bd0dffd781
 		fix(llm): preserve nested OpenAI stream errors (#36130) 
Co-authored-by: Harsh Mathur <jeevesforharsh@gmail.com>
Co-authored-by: Aiden Cline <aidenpcline@gmail.com>
 2026-07-09 14:40:28 -05:00
Shoubhit Dash
3cbcd8d727
 		feat(llm): complete provider package entrypoints (#35907) 2026-07-08 18:41:46 +05:30
opencode-agent[bot]
ed4f833813
 		fix(core): preserve provider metadata namespaces (#35817) 
Co-authored-by: Aiden Cline <63023139+rekram1-node@users.noreply.github.com>
Co-authored-by: Aiden Cline <aidenpcline@gmail.com>
 2026-07-07 21:45:58 -05:00
Aiden Cline
f488089179
 		fix: preserve compaction files and classify Z.AI overflow (#35747) 
Co-authored-by: opencode-agent[bot] <219766164+opencode-agent[bot]@users.noreply.github.com>
Co-authored-by: 冯基魁 <56265583+fengjikui@users.noreply.github.com>
 2026-07-07 09:54:31 -05:00
Kit Langton
bf01264661
 		feat(core): finalize session event lifecycle (#35272) 2026-07-06 18:14:06 -04:00
Aiden Cline
9e0d3976e1
 		chore: merge dev into v2 (#35591) 
Co-authored-by: Frank <frank@anoma.ly>
Co-authored-by: Aarav Sareen <96787824+arvsrn@users.noreply.github.com>
Co-authored-by: Brendan Allan <git@brendonovich.dev>
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
Co-authored-by: Jack <jack@anoma.ly>
Co-authored-by: Brendan Allan <14191578+Brendonovich@users.noreply.github.com>
Co-authored-by: Shoubhit Dash <shoubhit2005@gmail.com>
Co-authored-by: opencode-agent[bot] <219766164+opencode-agent[bot]@users.noreply.github.com>
Co-authored-by: James Long <longster@gmail.com>
Co-authored-by: Dustin Deus <deusdustin@gmail.com>
Co-authored-by: starptech <starptech@starptechs-MBP.fritz.box>
Co-authored-by: Luke Parker <10430890+Hona@users.noreply.github.com>
Co-authored-by: 𝓛𝓲𝓽𝓽𝓵𝓮 𝓕𝓻𝓪𝓷𝓴 <little-frank@opencord.local>
Co-authored-by: Dax <mail@thdxr.com>
Co-authored-by: usrnk1 <7547651+usrnk1@users.noreply.github.com>
Co-authored-by: Jay <53023+jayair@users.noreply.github.com>
Co-authored-by: runvip <164729189+runvip@users.noreply.github.com>
Co-authored-by: opencode <opencode@sst.dev>
Co-authored-by: Julian Coy <julian@ex-machina.co>
Co-authored-by: Vladimir Glafirov <vglafirov@gitlab.com>
Co-authored-by: Adam <2363879+adamdotdevin@users.noreply.github.com>
Co-authored-by: Kit Langton <kit.langton@gmail.com>
Co-authored-by: Simon Klee <hello@simonklee.dk>
Co-authored-by: Jay <air@live.ca>
Co-authored-by: David Hill <1879069+iamdavidhill@users.noreply.github.com>
 2026-07-06 16:05:29 -05:00
Shoubhit Dash
e57d9ca390
 		feat(core): flatten provider config and load native packages (#35563) 
Co-authored-by: Dax Raad <d@ironbay.co>
 2026-07-06 16:18:56 -04:00
James Long
4790a2772c
 		feat(simulation): add driver controlled backend LLM (#35186) 2026-07-03 14:19:25 -04:00
Aiden Cline
460cdc5aec
 		fix(llm): remove package retry policy (#35003) 2026-07-02 15:51:12 -05:00
Kit Langton
4045041554
 		refactor(core): simplify system context into a belief model (#34917) 2026-07-02 09:32:35 -04:00
Aiden Cline
140224b0fc
 		fix(core): derive models.dev reasoning variants (#34726) 2026-07-02 00:23:46 -05:00
James Long
8c94e9005f
 		chore: merge dev into v2 (#34788) 
Co-authored-by: Brendan Allan <14191578+Brendonovich@users.noreply.github.com>
Co-authored-by: Kit Langton <kit.langton@gmail.com>
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
Co-authored-by: Affan Ali <93028901+affanali2k3@users.noreply.github.com>
Co-authored-by: affanali2k3 <affanalikhanxx@gmail.com>
Co-authored-by: Frank <frank@anoma.ly>
Co-authored-by: opencode-agent[bot] <219766164+opencode-agent[bot]@users.noreply.github.com>
Co-authored-by: 𝓛𝓲𝓽𝓽𝓵𝓮 𝓕𝓻𝓪𝓷𝓴 <little-frank@opencord.local>
Co-authored-by: Aiden Cline <63023139+rekram1-node@users.noreply.github.com>
Co-authored-by: Jay V <air@live.ca>
Co-authored-by: Dax Raad <d@ironbay.co>
Co-authored-by: Aarav Sareen <96787824+arvsrn@users.noreply.github.com>
Co-authored-by: OpeOginni <107570612+OpeOginni@users.noreply.github.com>
Co-authored-by: Luke Parker <10430890+Hona@users.noreply.github.com>
Co-authored-by: Ben Guthrie <benjee.012@gmail.com>
Co-authored-by: Dax <mail@thdxr.com>
Co-authored-by: Filip <34747899+neriousy@users.noreply.github.com>
Co-authored-by: Max Anderson <max.a.anderson95@gmail.com>
Co-authored-by: Brendan Allan <git@brendonovich.dev>
Co-authored-by: Jack <jack@anoma.ly>
Co-authored-by: Shoubhit Dash <shoubhit2005@gmail.com>
Co-authored-by: Dustin Deus <deusdustin@gmail.com>
Co-authored-by: starptech <starptech@starptechs-MBP.fritz.box>
Co-authored-by: Aiden Cline <aidenpcline@gmail.com>
Co-authored-by: usrnk1 <7547651+usrnk1@users.noreply.github.com>
Co-authored-by: Jay <53023+jayair@users.noreply.github.com>
Co-authored-by: runvip <164729189+runvip@users.noreply.github.com>
Co-authored-by: opencode <opencode@sst.dev>
Co-authored-by: Julian Coy <julian@ex-machina.co>
Co-authored-by: Vladimir Glafirov <vglafirov@gitlab.com>
 2026-07-01 17:12:00 -04:00
Dax Raad
11d2f3e5f8 fix(llm): end reasoning before responses 2026-06-26 13:55:04 -04:00
Aiden Cline
f254476043
 		fix(llm): omit stateless response item ids (#34027) 2026-06-26 10:14:43 -05:00
Aiden Cline
5f61d21487
 		feat(llm): pass strict through tool definitions for Codex parity (#33392) 2026-06-25 16:28:39 -05:00
Kit Langton
516cfe4e09
 		refactor(schema): extract shared public schemas (#33571) 2026-06-23 22:31:06 -04:00
Kit Langton
130957288e
 		fix(llm): preserve structured tool errors (#33405) 2026-06-22 13:43:07 -04:00
Tommy D. Rossi
d5980b47e9
 		feat(llm): add video and audio media support to Gemini protocol (#31889) 
Co-authored-by: Aiden Cline <63023139+rekram1-node@users.noreply.github.com>
 2026-06-22 10:03:19 -05:00
Kit Langton
4f1ae1604e
 		chore: upgrade Effect to beta 83 (#32340) 2026-06-20 19:57:29 -04:00
Lucas Kim
2d993cd0d5
 		fix(experimental llm pkg): forward topK to Converse via additionalModelRequestFields (#33030) 2026-06-20 14:35:05 -05:00
opencode-agent[bot]
cc52dc396c chore: generate 2026-06-09 18:30:51 +00:00
Dax
132ef57272
 		refactor(core): simplify location filesystem (#31545) 2026-06-09 14:28:45 -04:00
Kit Langton
83dca45dd5
 		fix(core): make V2 reads media-aware and binary-safe (#31038) 2026-06-05 19:48:34 -04:00
Kit Langton
820c984d47
 		fix(core): recover v2 context overflow (#31005) 2026-06-05 16:05:51 -04:00
Kit Langton
0bdd9aa494
 		fix(core): preserve model request semantics (#30990) 2026-06-05 14:23:31 -04:00
Kit Langton
1af8dafd3e
 		feat(core): persist v2 session context epochs (#30789) 2026-06-04 23:26:43 -04:00
opencode-agent[bot]
134f4136da chore: generate 2026-06-04 23:58:11 +00:00
Kit Langton
f011d77128 fix(llm): normalize OpenAI function tool schemas 2026-06-04 19:56:31 -04:00
opencode-agent[bot]
b0a929440b chore: generate 2026-06-04 03:03:39 +00:00
Kit Langton
76ee87ead8
 		feat(core): add embedded v2 session runtime and tool foundation (#30632) 2026-06-03 23:02:17 -04:00
Dax
7f571d36ea
 		refactor(core): move database schema ownership (#29068) 
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
 2026-05-30 21:08:38 -04:00
opencode-agent[bot]
728eae9c25 chore: generate 2026-05-24 01:08:00 +00:00
Aiden Cline
eb84f461b8
 		fix(llm): split OpenAI reasoning summary blocks (#29000) 2026-05-23 20:06:45 -05:00
Aiden Cline
14c511e380
 		fix(llm): stabilize anthropic tool result typecheck (#28909) 2026-05-22 17:14:40 -05:00
Kit Langton
d0cb58782f
 		fix(llm): surface code, type, and nested fields on provider stream errors (#28757) 2026-05-22 12:37:55 -04:00
opencode-agent[bot]
8a55920531 chore: generate 2026-05-22 16:25:15 +00:00
Kit Langton
9db90a0b76
 		fix(llm): emit structured image blocks for tool-result media in Anthropic Messages (#28755) 2026-05-22 12:23:41 -04:00
Kit Langton
700d012025
 		fix(llm): emit structured input_image content for tool-result media in OpenAI Responses (#28754) 2026-05-22 12:23:23 -04:00
Kit Langton
e92c4fb460
 		chore: drop dead imports across opencode/core/llm (#28720) 2026-05-21 17:03:12 -04:00
Kit Langton
61390dbb49
 		fix(llm): preserve native continuation metadata (#28678) 2026-05-21 11:57:45 -04:00
Kit Langton
facd207396
 		fix(opencode): support native OpenAI OAuth fetch (#28571) 2026-05-20 23:29:25 -04:00
Kit Langton
16fb6dac8d
 		fix(llm): restore OpenAI reasoning streams (#28552) 2026-05-20 21:02:59 -04:00
Kit Langton
41f6daf96a
 		Refactor LLM route-first provider API (#28523) 2026-05-20 20:15:52 -04:00
Kit Langton
dbe36851bc
 		Preview native LLM runtime stack (#27114) 2026-05-18 14:41:36 -04:00
Kit Langton
ba5c8d3822
 		fix(llm): preserve tool error defects (#27403) 2026-05-13 20:43:32 -04:00
Kit Langton
dd14413a64
 		Preserve native LLM tool context (#27116) 2026-05-12 16:16:58 -04:00
Kit Langton
8030a6c187
 		Emit LLM stream lifecycle events (#26971) 2026-05-11 21:31:48 -04:00