feat(schemas): @tanstack/ai-schemas with nightly OpenAPI sync (closes #619)#622
Open
tombeckenham wants to merge 2 commits into
Open
feat(schemas): @tanstack/ai-schemas with nightly OpenAPI sync (closes #619)#622tombeckenham wants to merge 2 commits into
tombeckenham wants to merge 2 commits into
Conversation
Contributor
|
Important Review skippedToo many files! This PR contains 218 files, which is 68 over the limit of 150. To get a review, narrow the scope: ⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Run ID: ⛔ Files ignored due to path filters (1)
📒 Files selected for processing (218)
You can disable this status message by setting the Use the checkbox below for a quick retry:
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Contributor
🚀 Changeset Version Preview1 package(s) bumped directly, 0 bumped as dependents. 🟨 Minor bumps
|
|
Review the following changes in direct dependencies. Learn more about Socket for GitHub.
|
|
View your CI Pipeline Execution ↗ for commit ecc6fac
☁️ Nx Cloud last updated this comment at |
@tanstack/ai
@tanstack/ai-anthropic
@tanstack/ai-client
@tanstack/ai-code-mode
@tanstack/ai-code-mode-skills
@tanstack/ai-devtools-core
@tanstack/ai-elevenlabs
@tanstack/ai-event-client
@tanstack/ai-fal
@tanstack/ai-gemini
@tanstack/ai-grok
@tanstack/ai-groq
@tanstack/ai-isolate-cloudflare
@tanstack/ai-isolate-node
@tanstack/ai-isolate-quickjs
@tanstack/ai-mcp
@tanstack/ai-ollama
@tanstack/ai-openai
@tanstack/ai-openrouter
@tanstack/ai-preact
@tanstack/ai-react
@tanstack/ai-react-ui
@tanstack/ai-schemas
@tanstack/ai-solid
@tanstack/ai-solid-ui
@tanstack/ai-svelte
@tanstack/ai-utils
@tanstack/ai-vue
@tanstack/ai-vue-ui
@tanstack/openai-base
@tanstack/preact-ai-devtools
@tanstack/react-ai-devtools
@tanstack/solid-ai-devtools
commit: |
This was referenced May 25, 2026
tombeckenham
added a commit
that referenced
this pull request
May 26, 2026
Adds a fifth generic `TModelDurationByName` to `VideoAdapter` plus two
introspection methods on the base class:
- `availableDurations()` returns a `DurationOptions` tagged union
(`discrete | range | mixed | none`) describing the durations the
current model accepts.
- `snapDuration(seconds)` coerces raw seconds to the closest valid
duration for the current model.
`generateVideo({ duration })` is now typed via
`VideoDurationForAdapter<TAdapter>`. The FAL adapter derives its
per-model duration type from the SDK's `EndpointTypeMap`, so e.g.
`falVideo('fal-ai/kling-video/v1.6/standard/text-to-video')` types
`duration` as `'5' | '10'`; `falVideo('fal-ai/veo3')` types it as
`'4s' | '6s' | '8s'`; `falVideo('fal-ai/minimax/video-01')` rejects
the field entirely.
Adapters that have not yet declared their per-model duration map get
sensible defaults (`{ kind: 'none' }`, `undefined`) so existing video
adapters keep working without changes.
Built on top of #622 (`@tanstack/ai-schemas`); once that PR's FAL
pipeline syncs runtime constraint data, the hand-curated map in
`packages/typescript/ai-fal/src/video/video-provider-options.ts` can
be replaced with schema-derived lookups. Follow-up issue #634 covers
building the Gemini Veo adapter directly on this contract.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
6 tasks
…oses #619) JSON Schema + Zod definitions for AI provider generation endpoints, generated from each provider's official OpenAPI spec (or equivalent): OpenAI, Anthropic, Gemini, ElevenLabs, xAI Grok, and FAL (600+ models). Endpoints are grouped per provider by activity, aligned with the core library's activities — chat, image, video, audio (TTS, transcription, music, SFX in one group), embeddings, moderation — behind @tanstack/ai-schemas/{provider}/{activity}/{json-schema,zod} subpaths. Platform/admin endpoints (projects, invites, certificates, fine-tuning, file stores, workspaces) are excluded from generation. JSON Schemas ship self-contained with their $ref closures bundled under $defs; binary- response media endpoints map input-only entries. Pipeline (ported from fal-ai/fal-js#212, generalised multi-provider): fetch-schemas (per-provider fetchers) → @hey-api/openapi-ts (pinned, JSON Schema + Zod 4 plugins, per-activity spec splits with orphan pruning) → generate-endpoint-maps ($defs bundling, endpoint-id-keyed maps). Post-processing strips hey-api's schema-name discriminator grafts (which made discriminated unions reject every valid payload), accepts multipart/form-data bodies, and resolves dedup-renamed schema refs. .github/workflows/sync-schemas.yml refreshes specs daily and opens an automated PR on diff, following the sync-models.yml pattern. FAL specs are committed like the other providers'; the fetcher needs FAL_KEY (repo secret) and the plain `expand=openapi-3.0` query param. Squashed and rebased onto the flattened packages/* monorepo layout; the package lives at packages/ai-schemas. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
tombeckenham
force-pushed
the
619-create-schemas-library
branch
from
7a05fca to
d9904cf
Compare
Pulls the public OpenAPI 3.1 spec from openrouter.ai/openapi.json and classifies into chat (chat/completions, messages, responses), audio (speech, transcriptions), and video (/videos — including the frame_images / input_references image-conditioning shape). Account management, preset-scoped variants, and rerank drop out of generation. /embeddings is intentionally unclassified: OpenRouter declares its schemas inline rather than via $ref, so it can't be mapped to named exports. Per-model video constraints (supported_durations, resolutions, aspect ratios from GET /api/v1/videos/models) are a candidate follow-up — noted in the provider module. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
🎯 Changes
Closes #619.
Adds
@tanstack/ai-schemas— per-endpoint JSON Schema + Zod definitions for every supported provider's generation endpoints, generated nightly from upstream OpenAPI specs. Architecture ported from fal-ai/fal-js#212 and generalised to multi-provider.Why:
model-meta.tscovers coarse per-model facts (context window, modalities, pricing) but not the per-endpoint constraint surface — which durations Veo3 accepts, which image sizes Flux2 takes, valid ElevenLabs voice IDs, prompt-length caps. Schemas give runtime code and LLMs a way to shape and validate requests before they hit the network.Providers: OpenAI, Anthropic, Gemini (Discovery doc → OpenAPI in-pipeline), ElevenLabs, xAI Grok, OpenRouter (incl. async video generation with frame/reference image conditioning), FAL (600+ models, specs committed; CI fetch needs
FAL_KEY).Activity grouping: endpoints are grouped per provider by activity, aligned with the core library's activities —
chat,image,video,audio(TTS, transcription, music, SFX in one group),embeddings,moderation:Platform/admin endpoints (projects, invites, certificates, fine-tuning, file stores, workspaces) are excluded from generation. JSON Schemas ship self-contained (
$refclosures bundled under$defs); binary-response media endpoints (e.g. TTS) map input-only entries. No aggregator barrel — bundlers tree-shake by file.Pipeline:
fetch-schemas→@hey-api/openapi-ts(pinned 0.97.2, JSON Schema + Zod 4 plugins, per-activity spec splits with orphan pruning) →generate-endpoint-maps($defs bundling, endpoint-id-keyed maps). Post-processing fixes three generator/spec issues: hey-api's schema-name discriminator grafts (discriminated unions rejected every valid payload at runtime), multipart/form-data request bodies (transcription/dubbing endpoints were silently missing), and dedup-renamed schema refs dangling in$defs.Automation:
.github/workflows/sync-schemas.ymlruns daily at 05:00 UTC and opens an automated PR on diff, following thesync-models.ymlpattern. Repo setup: addFAL_KEYto repo secrets (the fetcher must use the plainexpand=openapi-3.0param — the bracketed form is silently ignored upstream).Coupled updates:
docs/advanced/ai-schemas.md(+docs/config.json), media-generation and adapter-configuration skills, changeset for the initial minor release.✅ Checklist
pnpm run test:pr.🚀 Release Impact
🤖 Generated with Claude Code