AI Chatbot

Project Overview

Full-stack AI chatbot starter with RAG, file uploads, and project-scoped chats, built on Next.js, Vercel AI SDK, Postgres, and Upstash Vector. It includes a production-ready architecture for auth, storage, vector search, usage controls, and deployment.

Key Features

• Streaming chat UI with Markdown rendering, auto-scroll, and rich AI elements.
• Multi-model support via Vercel AI Gateway (OpenAI, Anthropic, Google, xAI, and many more).
• File uploads + RAG with vector indexing and in-chat file search/read tools.
• Web search support for OpenAI, Anthropic, and Google Gemini models.
• Projects + chats with per-project history and organized sidebar.
• Auth + access control with public/private chat visibility.
• System emails for user account management.
• Usage tracking & limits with per-user chat credits.
• Advanced routing + client-side caching for seamless navigation and performance.

Architecture

• Backend:
  • AI layer: Vercel AI SDK + AI Gateway
  • DB layer: Neon Postgres + Drizzle ORM
  • RAG: Upstash Vector
  • File storage: Vercel Blob
  • Authentication: Better Auth
  • Emails: Resend API
• Frontend:
  • Next.js 16 + React 19 + React Query
  • shadcn/ui + Tailwind CSS
• Hosting:
  • Vercel Platform

Check out the project repository on Github for more info.