Skip to main content
FastAPI AI Kit
All use casesUse Case

A complete retrieval-augmented generation pipeline, pre-wired.

RAG involves more moving parts than most tutorials show: chunking strategy, embedding caching, vector store selection, context injection, and source attribution. FastAPI AI Kit ships all of it wired together, with pgvector and Qdrant supported out of the box.

FastAPIpgvectorQdrantOpenAI EmbeddingsPostgreSQLCelery

The usual pain points

  • ✕Building document ingestion, chunking, and embedding as separate one-off scripts
  • ✕Choosing between pgvector and Qdrant before understanding your scale needs
  • ✕Injecting retrieved context into prompts consistently across endpoints
  • ✕Returning source citations alongside generated answers

How the kit solves them

  • rag.ingest() and rag.query() cover chunking, embedding, storage, and retrieval in two calls
  • pgvector and Qdrant both supported — switch with one env var, no code changes
  • Context injection and prompt construction handled consistently by the RAG service
  • Query responses include source references by default, not as a bolt-on

Example implementation

main.py
await rag.ingest(source="handbook.pdf", collection="docs", chunk_size=512)

result = await rag.query(question="What's the PTO policy?", collection="docs", top_k=5)
# result.answer + result.sources, ready to return to the client

Ready to build your rag boilerplate?

FastAPI AI Kit ships with everything shown above, pre-configured and production-ready. Clone the repo and start building in minutes.

Ready to ship your AI backend this weekend?

Join developers who skipped weeks of boilerplate and went straight to building.

Read the docs
No subscriptions · One-time payment · Lifetime updates