Commit Graph

21 Commits

Author SHA1 Message Date
mushuiyu886 e8eb765ab7 fix(chutes): honor serving context limits (#119196) 2026-08-04 19:42:54 -07:00
Peter Steinberger 603f839058 refactor(providers): collapse live model discovery onto shared projection hook (#113903)
* feat(plugin-sdk): add live catalog row projection

* refactor(providers): share live catalog projection

* fix(venice): keep live projection internal
2026-07-25 16:17:27 -07:00
Peter Steinberger e30217d25c refactor(providers): shared manifest row builder and fleet default-model adoption (#113879)
* feat(plugin-sdk): add manifest model row builder

* refactor(providers): share manifest model row building

* refactor(providers): source defaults from manifests

* refactor(providers): drop unconsumed builder aliases

* fix(providers): keep Cohere catalog internal
2026-07-25 15:00:02 -07:00
Peter Steinberger be6ec97e11 refactor(providers): manifest-declared default models and unified onboarding presets (#113794)
* feat(model-catalog): declare provider default models

* refactor(providers): source onboarding defaults from manifests

* fix(cohere): remove unused model id export

* fix(chutes): preserve public default model id

* fix(model-catalog): accept provider default models
2026-07-25 12:54:53 -07:00
Peter Steinberger f703d803cc chore(models): curate all provider model catalogs to current-generation lineups (#113681)
* chore(models): fleet-wide provider catalog curation

* fix(models): sync provider runtime catalogs and tests with curated manifests

* test(models): align venice lifecycle assertions and copilot auth default with curated catalogs

* test(models): align catalog lifecycle contract checks
2026-07-25 07:21:17 -07:00
mushuiyu886 53f13f0510 fix(chutes): discover models when HTTP proxy is required (#111266)
Co-authored-by: Peter Steinberger <steipete@gmail.com>
2026-07-21 09:38:36 -07:00
mushuiyu886 367fe9a42b fix(chutes): report cache-read pricing from live catalog (#111253) 2026-07-19 19:56:42 -07:00
Peter Steinberger 349f78776d fix(models): refresh bundled provider catalogs (#109410)
* fix(models): refresh bundled provider catalogs

* docs(models): refresh generated docs map

* fix(xiaomi): keep provider helper private

* chore(release): defer catalog release note
2026-07-16 16:47:25 -07:00
Peter Steinberger d4183facf1 refactor(deadcode): tighten extension root modeling (#108538) 2026-07-15 18:18:23 -07:00
Jason (Json) 57e0bdaabe feat: add live provider model catalog helper
Summary:
- Add a shared live provider catalog runtime for SDK-backed providers.
- Route OpenAI, xAI, OpenCode Go, Chutes, DeepInfra, Venice, NVIDIA, and Vercel AI Gateway live model discovery through the shared helper.
- Remove duplicated provider-local live catalog caching and harden auth marker stripping, empty live-result retries, and OpenAI custom-base-url handling.

Verification:
- node scripts/run-vitest.mjs extensions/openai/openai-provider.test.ts src/plugin-sdk/provider-catalog-live-runtime.test.ts src/commands/models/list.source-plan.test.ts extensions/opencode-go/index.test.ts extensions/nvidia/provider-catalog.test.ts
- pnpm plugin-sdk:api:check
- pnpm lint --threads=8
- pnpm run lint:extensions:bundled
- pnpm run test:extensions:package-boundary:compile
- pnpm check:import-cycles
- pnpm exec oxfmt --check extensions/openai/openai-provider.ts extensions/openai/openai-provider.test.ts
- git diff --check origin/main...HEAD
- autoreview clean: no accepted/actionable findings reported
- AWS Crabbox focused remote proof: run_364680d1bff8 / cbx_2456fffafe01
- Earlier same-PR AWS Crabbox live proof: run_1f05ccab368e / cbx_7375c79fcf9b

Known proof gap:
- Final current-code true live-provider smoke was blocked by Crabbox secret hydration, documented in the PR proof comment.
2026-06-07 14:16:00 -07:00
Peter Steinberger a4a27517ff docs: document cerebras and chutes providers 2026-06-04 08:09:34 -04:00
Peter Steinberger ef7ad6f744 fix: validate chutes model token metadata 2026-05-28 18:01:07 -04:00
Peter Steinberger 4c210e22fa Adapt image compression quality by model (#85742)
* feat: adapt image compression quality

* refactor: move image limits into model metadata

* test: cover adaptive image downscaling

* test: cover image tool live providers

* fix: apply media metadata to all image paths

* fix: align providerless image compression

* fix: add chutes runtime image limits

* fix: optimize image data urls with model limits

* fix: type media metadata merge

* fix: optimize data url byte limits after decode

* fix: preserve data url optimizer fallback

* fix: keep low-side image compression fallbacks

* fix: enforce data url image compression policy

* fix: preserve gif data url media policy

* fix: satisfy adaptive image type checks

* test: keep cron provider-runtime mock current
2026-05-23 21:45:55 +01:00
Peter Steinberger 827b0de0ce refactor: reduce plugin sdk surface 2026-05-10 12:37:10 +01:00
Vincent Koc 4ae9ae12b6 test(e2e): fix kitchen sink crabbox coverage (#76287)
* test(e2e): fix kitchen sink crabbox coverage

* test(e2e): update kitchen sink expected diagnostics

* fix(plugins): harden registry and package gates

* fix(plugins): load lazy tool middleware snapshots

* fix(ci): satisfy crabbox branch gates

* fix(plugins): await guarded fetch cleanup
2026-05-02 18:39:26 -07:00
Vincent Koc 52cd76a9e2 fix(providers): isolate model discovery test env guards 2026-05-02 17:35:39 -07:00
Peter Steinberger 5fa3b8d7a0 refactor: dedupe provider trimmed readers 2026-04-08 00:09:41 +01:00
Peter Steinberger a4bb2698dd refactor: dedupe ui provider lowercase helpers 2026-04-07 20:58:01 +01:00
Peter Steinberger 5489bff7c3 test: speed up chutes model tests 2026-04-07 11:35:58 +01:00
Peter Steinberger b39a7e8073 fix: break plugin-sdk provider barrel recursion 2026-03-27 23:47:04 +00:00
Peter Steinberger c28e76c490 refactor: move provider model helpers into plugins 2026-03-27 20:20:31 +00:00