The right AI library
for every JavaScript app
you build with your agent

The TypeScript toolkit for building AI apps. Streaming UI, tool calling, and multi-provider support.

Type-safe, provider-agnostic TypeScript AI SDK from TanStack: streaming chat, tool calling, agents, structured output and multimodal apps, with React, Vue, Svelte and Solid bindings.

Low-level orchestration for stateful agents as graphs. Durable execution, human-in-the-loop, and memory.

Framework for building LLM-powered apps. Chains, agents, retrieval, and memory out of the box.

Open-source TypeScript framework for AI agents and workflows. Memory, observability, and integrations with React, Next.js, and Node.

Lightweight framework for multi-agent workflows: agents with tools, handoffs, guardrails, sessions, and tracing. Provider-agnostic.

The agent harness behind Claude Code as a library: an agent loop with file, shell, and web tools, subagents, and MCP.

Cloudflare's SDK for AI agents that run on its network, each with its own state, storage, scheduling and WebSocket connections (built on Durable Objects).

Vercel's TypeScript toolkit for chat bots and AI agents on Slack, Microsoft Teams, Google Chat, Discord, WhatsApp and more, from one codebase.

Deprecated by LlamaIndex and no longer maintained. Listed for reference; for RAG in TypeScript use another framework here with a vector database.

Archived TypeScript library for multimodal AI apps. Retained as a historical reference; use AI SDK for new projects.

Official TypeScript SDK for Claude. Build conversational AI, code generation, and analysis tools.

Official Node.js library for the OpenAI API. GPT, DALL-E, Whisper, and embeddings.

Official JavaScript and TypeScript SDK for Gemini. Use the maintained @google/genai package.

Official TypeScript SDK for Mistral AI models. Fast, efficient open-weight model access.

Official TypeScript SDK for OpenRouter: one API key and one API for hundreds of models from many providers.

Official TypeScript SDK for Cohere. Embeddings, reranking, and RAG-optimized models.

Official SDK for the Model Context Protocol: build MCP servers that expose tools and data to AI apps, and clients that connect to them.

Durable execution, events, and queues: step functions that retry and resume, cron jobs, and background work on any serverless platform.

Open-source platform for long-running AI tasks and agents in TypeScript: no timeouts, retries, queues, human-in-the-loop waits, and realtime streaming.

Durable workflows as code on Temporal: workflows survive crashes and restarts and can run for days.

Machine learning for JavaScript. Train and deploy ML models in the browser and Node.js.

Run ONNX models in the browser with WebAssembly and WebGPU acceleration.

Run Hugging Face transformers directly in the browser. NLP, vision, and audio — no server needed.

Run large language models entirely in the browser using WebGPU. Private, fast, offline-capable.

GPU-accelerated neural networks in JavaScript. Simple API for training and running neural nets.

Embedded vector database that runs inside your app and stores data on local disk or object storage. Vector, full-text, and SQL search.

Small vector search extension for SQLite. Runs anywhere SQLite runs, including the browser with WebAssembly.

TypeScript client for Pinecone vector database. Serverless vector search at scale.

JavaScript client for Chroma, the open-source embedding database for AI apps.

TypeScript client for Weaviate vector database. Semantic search and hybrid retrieval.

TypeScript client for Turbopuffer. Fast, cost-effective serverless vector database.

Anthropic's agentic coding tool. An AI assistant that lives in your terminal and understands your codebase.

Minimal, extensible open-source coding agent for the terminal. Its packages also work on their own: a unified LLM API (pi-ai) and an agent core (pi-agent-core).

AI-first code editor built on VS Code. Inline AI editing, chat, and codebase-aware completions.

AI pair programmer by GitHub. Code suggestions, chat, and pull request summaries.

AI-powered UI generation. Describe a component and get production-ready React code.