AI product Open source
A minimal chatbot starter built with Next.js, the AI SDK, shadcn/ui, shadcn/react, shadcn/typeset, and the Vercel AI Gateway. It provides streaming chat with Markdown rendering, provider-native web search, tool calling, and a human-in-the-loop questionnaire in which the model can ask clarifying questions.
The chat route streams responses with `streamText`, while the interface renders assistant messages as typed parts for text, GitHub repository lookups, web searches, questionnaires, and source URLs. Tool definitions are kept in separate files and composed through a central index; tool-part components display progress, results, errors, answered questions, and deduplicated search citations.
Vercel deployments authenticate to the AI Gateway through OIDC, while local development uses an AI Gateway API key. The public unauthenticated chat route includes request validation, model restrictions, output and step limits, and client-disconnect handling, but the README recommends adding rate limiting, spending limits, and authentication before exposing it to real traffic.
3 uses taken from transcripts — each links to the moment in the video.
A minimal starting point for building a streaming AI chat app in Next.js, with Markdown rendering, tool calling, and human-in-the-loop questionnaires. Its typed tools connect inferred input and output types to the UI.
A minimal open-source chatbot starter from the shadcn team, built with Next.js and the AI SDK. It includes streaming chat, Markdown rendering, tool calling, provider-native web search, clarification questions, typed streaming message parts, and one-click deployment.
A minimal Next.js and AI SDK starter for building chatbot interfaces. It includes token streaming, tool calling, web search, and human-in-the-loop questions, with one-click deployment to Vercel.
3 in the library.