Skip to content

CHANGELOG.md

Latest commit

 

History

History
86 lines (59 loc) · 7.98 KB

CHANGELOG.md

File metadata and controls

86 lines (59 loc) · 7.98 KB

@cloudflare/tanstack-ai

0.1.4

Patch Changes

  • #448 054ccb8 Thanks @threepointone! - Fix image inputs for vision-capable chat models
    • Handle all LanguageModelV3DataContent variants (Uint8Array, base64 string, data URL) instead of only Uint8Array
    • Send images as OpenAI-compatible image_url content parts inline in messages, enabling vision for models like Llama 4 Scout and Kimi K2.5
    • Works with both the binding and REST API paths

0.1.3

Patch Changes

  • #435 7381171 Thanks @mdhruvil! - Fix workers-ai adapter silently dropping image content parts.

  • #424 b2eeca8 Thanks @vaibhavshn! - Avoid duplicate tool call IDs by generating unique IDs per tool call index instead of trusting backend-provided IDs

  • #411 af08464 Thanks @baldyeagle! - Annotate createAnthropicChat to improve client type narrowing

  • #398 40e53c8 Thanks @vaibhavshn! - fix: add run/ prefix to workers-ai gateway endpoint and make API key optional for gateway bindings

  • #444 414b4d5 Thanks @mchenco! - Add sessionAffinity option to WorkersAiAdapterConfig for prefix-cache optimization. Routes requests with the same key to the same backend replica via the x-session-affinity header. Supported across binding, REST, and gateway modes.

0.1.2

Patch Changes

  • #406 9af703b Thanks @vaibhavshn! - Pass API Key properly for Gemini Tanstack AI Adapter

  • #400 8822603 Thanks @threepointone! - Add config validation to all Workers AI adapter constructors that throws a clear error when neither a binding, credentials (accountId + apiKey), nor a gateway configuration is provided. Widen all model type parameters (WorkersAiTextModel, WorkersAiImageModel, WorkersAiEmbeddingModel, WorkersAiTranscriptionModel, WorkersAiTTSModel, WorkersAiSummarizeModel) to accept arbitrary strings while preserving autocomplete for known models.

0.1.1

Patch Changes

  • #396 2fb3ca8 Thanks @threepointone! - - Update model recommendations: Aura-2 EN for TTS, Llama 4 Scout for chat examples
    • Add Aura-2 EN/ES to TTS model type
    • Preserve image/vision content in user messages instead of stripping to text-only
    • Add non-streaming fallback when REST streaming fails (GPT-OSS, Kimi)
    • Warn on premature stream termination instead of silently reporting "stop"
    • Consistent console.warn prefix for SSE parse errors
    • Move @cloudflare/workers-types from optionalDependencies to devDependencies (types-only, no runtime use)
    • Fix @openrouter/sdk version mismatch type errors

0.1.0

Minor Changes

  • #389 a4b756e Thanks @vaibhavshn! - Add @cloudflare/tanstack-ai — adapters for using TanStack AI with Cloudflare Workers AI and AI Gateway.

    Workers AI adapters

    All Workers AI adapters support four configuration modes: plain binding (env.AI), plain REST (account ID + API key), AI Gateway binding (env.AI.gateway(id)), and AI Gateway REST (account ID + gateway ID).

    • Chat (createWorkersAiChat) — Streaming chat completions via the OpenAI-compatible API. Includes tool calling with full round-trip support, structured output via json_schema, and reasoning text streaming (STEP_STARTED/STEP_FINISHED AG-UI events) for models like QwQ, DeepSeek R1, and Kimi K2.5. A custom fetch shim translates OpenAI SDK calls to env.AI.run() for binding mode, with a stream transformer that handles both Workers AI native format and OpenAI-compatible format.
    • Image generation (createWorkersAiImage) — Stable Diffusion and other text-to-image models.
    • Transcription (createWorkersAiTranscription) — Speech-to-text via Whisper and Deepgram Nova-3.
    • Text-to-speech (createWorkersAiTts) — Audio generation via Deepgram Aura-1.
    • Summarization (createWorkersAiSummarize) — Text summarization via BART-large-CNN.
    • Embeddings (createWorkersAiEmbedding) — Text embeddings (implemented but not yet exported, pending TanStack AI's BaseEmbeddingAdapter).

    AI Gateway adapters (third-party providers)

    Route requests through Cloudflare AI Gateway for caching, rate limiting, and unified billing. Each adapter injects a custom fetch (or httpOptions for Gemini) that handles both binding and credential-based gateway configurations.

    • OpenAI — Chat, summarize, image, transcription, TTS, video (createOpenAi*)
    • Anthropic — Chat, summarize (createAnthropic*)
    • Gemini — Chat, summarize, image, TTS (createGemini*). Credentials-only (Google GenAI SDK lacks custom fetch support).
    • Grok — Chat, summarize, image (createGrok*)
    • OpenRouter — Chat, summarize, image (createOpenRouter*). Accepts any model string.

    Utilities

    • createGatewayFetch — Shared fetch factory that routes requests through AI Gateway (binding or REST), with support for cache control headers (skipCache, cacheTtl, customCacheKey, metadata).
    • createWorkersAiBindingFetch — Fetch shim that makes env.AI look like an OpenAI endpoint, including stream transformation and tool call ID sanitization for the binding's strict [a-zA-Z0-9]{9} validation.
    • Config detection helpers (isDirectBindingConfig, isDirectCredentialsConfig, isGatewayConfig) using structural typing to discriminate env.AI from env.AI.gateway(id).
    • Shared binary utilities for normalizing Workers AI responses (Uint8Array, ArrayBuffer, ReadableStream, JSON wrapper) to base64.

    Robustness

    • Premature stream termination detection — if Workers AI truncates a response or the connection drops (no finish_reason), the adapter emits proper closing events so consumers don't hang.
    • Graceful non-streaming fallback — if a model returns a complete response despite stream: true, the binding shim wraps it into a valid response.
    • Deepgram Nova-3 transcription uses raw binary audio via REST (not JSON), automatically detected by model name.

    Testing

    • Comprehensive unit tests (186 tests) covering all adapters, config modes, stream transformation, message building, tool calling, reasoning events, premature termination, and public API surface.
    • E2E integration tests against real Workers AI APIs (both binding and REST paths) across 12 chat models + 4 transcription models + image/TTS/summarize, validating chat, multi-turn, tool calling, tool round-trips, structured output, reasoning, and all non-chat capabilities.
    • Tree-shakeable package exports with per-adapter entry points for ESM and CJS.