Next.js · Supabase pgvector · Claude

A RAG assistant over your docs —
with citations and proof.

Production-ready starter: streaming answers with cited sources, pgvector + RLS ready, and an eval harness to prove it doesn't hallucinate. One-time payment, self-hosted, 100% yours.

Buy — $129 Try the live demo no signup · no API keys
 eval/run.ts
$ npx tsx eval/run.ts PASS pricing.question grounded cited: pricing.md PASS support.refund grounded cited: support.md PASS product.limits grounded cited: product.md ────────────────────────────────────────── 3/3 passed · 0 hallucinations · avg 240ms
The differentiator

Other starters give you a chatbot.
We give you a way to prove it works.

CiteKit ships with an eval harness — contract, judge, and cases: define questions and expected answers, and measure whether your assistant grounds its replies in your documents. It's the difference between "looks like it works" and "I can prove it works." No other RAG boilerplate ships this out of the box.

What's inside

The full pipeline. No lock-in.

RAG with citations

Streaming chat with expandable sources. Vector retrieval via pgvector, lexical fallback in demo mode.

Supabase + RLS

Migration ready: documents table, ivfflat index, match_documents() and RLS policies.

Eval harness

Contract + judge + cases to measure quality on your own domain. The differentiator.

No SDKs

Claude + OpenAI-compatible embeddings via fetch. Swap providers without rewriting anything.

Ingest included

Script that chunks, embeds and inserts your .md files. Drop in your docs and go.

Demo without keys

Green build and working preview without Supabase or Claude. Try it in a minute.

How it works

Three steps to deploy.

Index

Drop your documents in content/ and run the ingest script.

Connect

Supabase + your Claude API key. Or try it in demo mode without any keys.

Deploy and measure

Deploy to Vercel (standalone) and run the eval to validate quality.

Pricing

One-time payment. Self-hosted. 100% yours.

$199$129
launch price
  • Full repo + updates
  • Eval harness for your domain
  • "Deploy in 10 minutes" guide
  • Unlimited projects (commercial license)
  • Guarantee: if the demo doesn't run, we refund you
Buy — $129
FAQ

What you'll want to know.

Do I need API keys to try it?

No. Demo mode runs without Supabase or Claude: lexical retrieval over seed data and answers assembled from the retrieved fragments.

Which models does it support?

Claude by default (configurable) and embeddings compatible with the OpenAI API. Everything via fetch, no SDKs.

Can I use it for client projects?

Yes, unlimited personal or client projects. You cannot resell the starter itself. See the license.

Why buy instead of building it myself?

It saves you the full pipeline — embeddings, pgvector, retrieval, streaming, citations, RLS — plus the evaluation layer, which almost nobody ships. Pay once and self-host.