Readable over magical
Every API route, prompt and retrieval setting is plain JavaScript in your repo. If you cannot read it, we did not ship it.
About FastRAG
Skip the plumbing. Ship the product.
FastRAG is a Next.js starter kit that packages the repetitive parts of retrieval-augmented generation — ingestion, embeddings, vector search and streaming chat — so you can spend your time on what makes your app different.
Principles
What FastRAG believes
Pay once, own it
A single purchase with lifetime updates from the private repository. No subscription and no per-seat pricing.
Swappable parts
The LLM, embeddings and vector store are each isolated in one file under lib/, so you can swap providers without a rewrite.
The stack, in full
No hidden services. These are the pieces the kit is built on.
- Framework
- Next.js (Pages Router), React, Tailwind CSS
- Chat model
- Anthropic Claude Haiku, streamed over SSE
- Embeddings
- Voyage AI voyage-3.5
- Vector store
- Pinecone serverless, one namespace per tenant
- Ingestion
- Cheerio for URLs, pdf-parse + Formidable for PDFs
- Orchestration
- LangChain.js text splitting
Built by one developer, for developers
FastRAG is built and maintained independently. Customers get access to a private GitHub repository, lifetime updates and a direct line to the person who wrote the code — expect a reply within 24 hours on weekdays.
Ready to build AI that knows your data?
Grab the source, wire up your API keys, and have a production RAG app running today.
Pay once · MIT license · Lifetime updates