Chatbot by Vercel

Open-source Next.js chatbot template built on the Vercel AI SDK

Free· Free and open-source (MIT-style). Users cover their own hosting costs (e.g. Vercel plan) and AI provider API usage.CodingProvider-agnostic via the Vercel AI SDK (OpenAI, Anthropic, Google, xAI, Groq, Mistral, and others configurable)

Visit website →

Best for

Product and platform engineers who want a modern, provider-agnostic Next.js codebase to fork as the foundation of an internal or customer-facing chatbot, rather than assembling streaming, auth, and tool-calling from scratch.

Skip if

Non-technical users, teams looking for a hosted no-code chatbot builder, or anyone who wants a turnkey SaaS with billing, analytics, and support included out of the box.

Chatbot by Vercel is a free, open-source Next.js template for building production-ready conversational AI applications. Distributed through the Vercel AI SDK ecosystem (ai-sdk.dev) and its companion GitHub repository (github.com/vercel/chatbot), it gives developers a batteries-included starting point rather than a hosted SaaS product. Out of the box the template ships with modern chat UX patterns such as streaming responses, message parts, resumable streams, reasoning traces, and artifacts (a side-panel for code, documents, or spreadsheets the model produces), along with theming, font customization, and tool-approval workflows for gating agentic actions. Because it is built on the Vercel AI SDK, the underlying model is pluggable: developers can wire it to OpenAI, Anthropic, Google, xAI, Groq, Mistral, or self-hosted providers by swapping a few lines of provider configuration, and can add tools, RAG retrieval, and auth on top. Typical audiences include product engineers who need an internal Copilot-style assistant, startups prototyping a customer-facing chatbot, and hobbyists who want a serious codebase to fork instead of assembling one from scratch. Common workflows are: deploy to Vercel with a one-click template, add an API key for the chosen LLM provider, then extend it with domain-specific tools, retrieval, and branding. It is not a no-code product and it is not a hosted chatbot service, but for teams comfortable in TypeScript it removes weeks of scaffolding around auth, persistence, streaming, and multi-modal input.

Editor's take

This is the reference chatbot codebase a lot of teams end up cloning anyway, just officially maintained. If you are building on the Vercel AI SDK it is close to a no-brainer starting point; the artifacts panel, resumable streams, and tool-approval scaffolding are exactly the pieces most teams underinvest in. Just remember it is a template you have to operate, not a product you buy.

— The AI Tool Bible editorial team

Pros

✅ Free and open-source under a permissive license, with an active repo maintained by Vercel
✅ Provider-agnostic through the Vercel AI SDK, so you can swap between OpenAI, Anthropic, Google, xAI, Groq, and others
✅ Ships with modern UX patterns most teams reinvent: streaming, reasoning traces, artifacts panel, resumable streams, message parts
✅ One-click deploy to Vercel plus a clean Next.js App Router codebase that is straightforward to fork and extend
✅ Includes tool-approval workflows and theming hooks, so agentic actions and branding are first-class rather than bolt-ons
✅ Serves as the reference implementation for the AI SDK, so it tracks new SDK features quickly

Cons

⚠️ It is a template, not a hosted product; you own deployment, scaling, auth, and observability
⚠️ Requires TypeScript, Next.js, and general web-app operations knowledge to customize seriously
⚠️ You pay for underlying model API usage separately, and costs can surprise teams that ship without guardrails
⚠️ Opinionated toward the Vercel deployment target; running it elsewhere works but loses some conveniences
⚠️ Documentation lives across the repo, AI SDK docs, and blog posts rather than in one polished product manual

Use cases

Internal engineering copilotCustomer-facing support chatbotRAG-backed knowledge assistantAgentic tool-using assistantMulti-provider LLM evaluation harnessPrototype for a startup MVP chatbotReference implementation for the Vercel AI SDKWhite-labelled branded assistant