Coming liveWaitlist open

Stop guessing why your RAG answer is wrong. Get the root cause in 2 seconds.

RAG Doctor. Tells you why your RAG failed, in 2 seconds.

Join the waitlist
Source on GitHub ↗
rag-doctor
RD
Why it exists

01PROBLEM

RAG pipelines fail silently: fragmented chunks, bad retrieval, or the right document in the wrong position.

02SOLUTION

A Python diagnosis tool. Pass the query, answer and retrieved docs, and it returns the root cause, severity and a config fix.

03VALUE

Debug RAG systematically instead of tweaking prompts at random.

Screenshots

The story

How to use it

  1. Install it with pip install rag-doctor.
  2. Pass your query, the answer, the retrieved documents and the expected answer.
  3. Read the root cause, its severity and a suggested configuration fix.

How it's built

A Python library with a set of diagnostic checks for common RAG failures (chunk fragmentation, retrieval misses, context position bias) that returns a structured report.

Why it matters for businesses

RAG systems fail silently and teams burn time guessing. A root cause in seconds turns debugging into a checklist and protects trust in AI answers.

In real life

A support bot keeps citing the wrong policy. RAG Doctor shows the right document was retrieved but buried, and suggests turning on a reranker.

Under the hood

How it's built.

// rag-doctor · AI tools
export const ragDoctor = {
stack: ["Python", "pip install rag-doctor"],
audience: ["AI engineers", "Teams running RAG in production"],
status: "coming", price: 0
}
Shape what's next

Feature requests.

No requests for RAG Doctor yet. Be the first!

Chat with Bala