Full-Stack AI Agent Template
A FastAPI + Next.js scaffold for building production AI agent products
- Best for
- Teams building a conversational AI or agent product who want a FastAPI/Next.js foundation with multiple LLM providers and RAG already wired up.
- Skip if
- You want a lightweight chat starter - this is a large, feature-dense template with a learning curve.
A full-stack starter for AI/LLM applications pairing a FastAPI + async SQLAlchemy backend with a Next.js 15/React 19 frontend. It supports multiple agent frameworks (PydanticAI, LangChain, LangGraph) and LLM providers (OpenAI, Anthropic, Gemini, OpenRouter), with RAG via Milvus/Qdrant/ChromaDB/pgvector, WebSocket streaming with tool visualization, conversation persistence, Stripe billing, and an admin panel.
Pros
- Supports multiple agent frameworks and LLM providers out of the box, not locked to one vendor
- RAG-ready with a choice of vector stores
- WebSocket streaming with real-time tool-call visualization
- Ships with CLAUDE.md for AI-agent-assisted development
Cons
- Large surface area (5 agent frameworks, 4 background task options) can be overwhelming to configure
- MIT licensed but maintained by a consultancy - long-term update cadence unverified
- Requires running both a Python backend and Next.js frontend, more moving parts than a single-stack starter
Features
Auth and access
- Auth included
Payments and billing
- Payments included
- Subscriptions
AI capability
- AI chat
- AI agents
- RAG / vector search
- Streaming responses
- BYO API key
- Model-agnostic
AI-friendliness
- CLAUDE.md
What's in the box
- Admin dashboard
- Outbound webhooks
Alternatives
SvelteKit AI Chatbot (jianyuan)
An open-source SvelteKit port of Vercel's AI Chatbot template
Laravel AI Chat
Laravel + Inertia AI chat starter with streaming responses across 9 model providers
Cloudflare Agents Starter
An AI chat agent starter on Cloudflare Workers with Auth0 authentication