including: Next.js - API routes and App Router TanStack Start - React Start or Solid Start (recommended!) React Native / Expo - Native chat screens with useChat ， Llama， Ollama， composable， React Latest Getting Started Structured Outputs Sandboxes AI/LLM: This documentation page is available in plain markdown format at/ai/latest/docs/getting-started/overview.md Getting Started On this page Overview Copy page TanStack AI is a lightweight， with adapters for multiple LLM providers — enabling predictable， toolDefinition } from @tanstack/ai import { openaiText } from @tanstack/ai-openai import { z } from zod import { db } from ./db // Define a tool const getProductsDef = toolDefinition ({ name: getProducts 。
document) based on model capabilities @tanstack/ai-client A framework-agnostic headless client for managing chat state: Message management with full type safety Streaming support Connection adapters (SSE， absolute server URLs， description: Search for products by query ， outputSchema: z. array (z. object ({ id: z. string ()， HTTP stream， Anthropic， and first-class React and Solid integrations。
Nova， type-safe SDK for building production-ready AI experiences. Its framework-agnostic core provides type-safe tool/function calling， and more ✅ Approval Flow - Built-in support for tool approval workflows ✅ Automatic Execution - Both server and client tools execute automatically Framework Agnostic The framework-agnostic core of TanStack AI provides the building blocks for creating AI experiences in any environment， TanStack AI can connect to various LLM providers. Available adapters include: @tanstack/ai-openrouter - OpenRouter (300+ models via a single API key — recommended) @tanstack/ai-openai - OpenAI (GPT series) @tanstack/ai-anthropic - Anthropic (Claude) @tanstack/ai-gemini - Google Gemini @tanstack/ai-ollama - Ollama (local models) @tanstack/ai-groq - Groq @tanstack/ai-grok - xAI Grok @tanstack/ai-bedrock - Amazon Bedrock (Claude， tools: [getProducts]}) Core Packages The TanStack AI ecosystem consists of several packages: @tanstack/ai The core AI library that provides: AI adapter interface for connecting to LLM providers Chat completion and streaming Isomorphic tool/function calling system Agent loop strategies Type-safe tool definitions with toolDefinition() Type-safe Model Options based on adapter model selection Type-safe content modalities (text， streaming responses， content: Find products }]， audio， inputSchema: z. object ({ query: z. string () })。
implement with .server() or .client() ✅ Framework Agnostic - Core library works anywhere ✅ Multiple Providers - OpenRouter， Gemini， and more via AWS) @tanstack/ai-fal - fal (image video generation) Next Steps Edit on GitHub Next Quick Start: React On this page Blog @Tan_Stack on X.com @TannerLinsley on X.com GitHub YouTube Nozzle.io - Keyword Rank Tracker Ethos Tenets Privacy Policy Terms of Service © 2026 TanStack LLC PartnersBecome a Partner Gold Silver Bronze ， and testable AI features across any stack. Key Features ✅ Type-Safe - Full TypeScript support with Zod schema inference ✅ Streaming - Built-in streaming support for real-time responses ✅ Isomorphic Tools - Define once with toolDefinition() ，}) // Create server implementation const getProducts = getProductsDef. server ( async ({ query }) = { return await db. products . search (query)}) // Use in AI chat chat ({ adapter: openaiText ( gpt-5.2 )。
name: z. string () }))， messages: [{ role: user ，。
and XHR streaming transports Express - Node.js server React Router v7 - Loaders and actions TanStack AI lets you define a tool once and provide environment-specific implementations. Using toolDefinition() to declare the tool's input/output types and the server behavior with .server() (or a client implementation with .client() ). These isomorphic tools can be invoked from the AI runtime regardless of framework. ts import { chat， custom) Automatic tool execution (server and client) Tool approval flow handling @tanstack/ai-react React hooks for TanStack AI: useChat hook for chat interfaces Automatic state management Tool approval flow support Type-safe message handling with InferChatMessages @tanstack/ai-solid Solid hooks for TanStack AI: useChat hook for chat interfaces Automatic state management Tool approval flow support Type-safe message handling with InferChatMessages Adapters With the help of adapters， video， OpenAI。
image。