The AI Toolkit for TypeScript. From the creators of Next.js, the AI SDK is a free open-source library for building AI-powered applications and agents
The cleanest way to wire streaming LLM responses into a React/Next.js UI - handles the hard parts (streaming, tool calls, edge runtimes) so you don't have to.
Where ai fits, based on its real tags and topics.
git clone https://github.com/vercel/ai.gitSee the repository's README for language-specific setup steps.
The AI SDK is a provider-agnostic TypeScript toolkit designed to help you build AI-powered applications and agents using popular UI frameworks like Next.js, React, Svelte, Vue, Angular, and runtimes like Node.js.
To learn more about how to use the AI SDK, check out our API Reference and Documentation.
You will need Node.js 22+ and npm (or another package manager) installed on your local development machine.
25.8K
Stars
4.9K
Forks
144
Watchers
1.7K
Open Issues
683
Contributors
15
Topics
Category: 🤖 AI Agents
Difficulty: Intermediate
Part of: AI Agents Toolkit
NVIDIA and Japan Bring Full-Stack AI and Robotics to Every Industry
Jul 15, 2026
China’s Moonshot AI releases Kimi K3, the largest open-source model ever, rivaling top U.S. systems
Jul 16, 2026
Nimble claims its new, domain-specialized Web Search Agents cut token costs in half while boosting retrieval accuracy
Jul 29, 2026
“The most approachable of the major frontend frameworks - official docs and tooling are exceptionally well designed for newcomers.”
“The default choice for production React today - server components, file-based routing, and first-class deployment story make it the fastest path from idea to shipped app.”
The Postgres development platform. Supabase gives you a dedicated Postgres database to build your web, mobile, and AI applications.
“A genuinely open-source Firebase alternative - real Postgres underneath, with auth, storage, and realtime built in rather than a proprietary database.”
“Still the most widely deployed UI library in the industry - the ecosystem, tooling, and hiring pool around it are unmatched.”
“The most widely adopted framework for chaining LLM calls, tools, and retrieval into real applications - a near-default starting point for RAG and agent projects.”
“The official, minimal-overhead client for OpenAI's API - the right starting point before reaching for a bigger framework.”
“The standard library for working with open-weight models - thousands of pretrained models are one line of code away.”
“One of the earliest serious multi-agent orchestration frameworks - useful once a single agent/chain isn't enough for the problem.”
“A lighter-weight take on multi-agent orchestration - agents get explicit roles and goals, which reads more naturally than raw chain composition.”
“Unifies dozens of LLM providers behind a single API, making model switching almost effortless in production.”
“The industry-standard container orchestrator - not always necessary at small scale, but the default assumption for any serious cloud-native platform team.”
“The most capable open-source relational database available - strong standards compliance, extensions for nearly everything (including full-text search and vectors), and a default choice for good reason.”
“The modern default frontend build tool - native ESM in dev means a dev server that starts and reloads almost instantly, even on large apps.”
“The open-source engine behind Docker - understanding containers here pays off no matter which orchestrator sits on top later.”
“The standard tool for infrastructure-as-code across every major cloud provider - one declarative language instead of provider-specific consoles or scripts.”
“Brings analytical SQL performance to local development without requiring a separate database server.”
