Hybrid retrieval that finds the number
Dense + sparse (BGE-M3) RRF fusion surfaces exact table figures that keyword- or vector-only search misses.
Trusted answers from your documents
Stop searching documents.Start getting answers.
docuQ turns your knowledge bases into a precise question-answering engine — hybrid retrieval, exact figures, every answer traced to source.
Built for teams that live in documents
Your teamFinanceLegalResearchOpsSupport
Retrieval built to find the exact answer.
Knowledge bases, your way
Per-user KBs, private by default, shareable read-only. Query across several knowledge bases at once.
Answers you can trust
Grounded generation with citations — every figure traces back to a source chunk you can open.
A real developer API
Query your KBs over HTTP with an API key. Embed docuQ inside your own product.
Three steps to a cited answer.
- 01
Create a knowledge base
Private by default. Spin up as many as you need, scoped to you or your team.
- 02
Add documents
Drop in PDFs or markdown. docuQ chunks, embeds, and indexes them for retrieval.
- 03
Ask in natural language
Ask a question and get a precise, cited answer traced straight to the source.
For developers
Query your knowledge bases over HTTP.
A real REST API behind an API key. Ask a question, name the knowledge bases, get a grounded answer with citations — ready to drop into your own product.
Read the API referencecurl -X POST https://your-docuq/api/v1/query \
-H "Authorization: Bearer dq_live_…" \
-H "Content-Type: application/json" \
-d '{"query":"lower-quartile CEO total pay?","knowledge_base_ids":["…"]}'Simple pricing. Start free.
Free Trial
Free· 7-day trial, no card required
- 1 knowledge base
- 50 document pages / month
- 100 queries / month
Pro
₹799/mo + GST
- 5 knowledge bases
- 1,000 document pages / month
- 2,000 queries / month
Max
₹2499/mo + GST
- 25 knowledge bases
- 5,000 document pages / month
- 8,000 queries / month
Start with a 7-day free trial — no card required. Upgrade when your team is ready, cancel anytime. Prices shown exclude GST, added at payment.
Technical foundation
Serious retrieval, yours to own.
- DeepSeek-chat answers
- BGE-M3 dense + sparse embeddings
- Modal GPU
- pgvector hybrid retrieval
- Self-hostable
Questions, answered.
What is docuQ?
docuQ turns your documents into a private question-answering engine. Upload PDFs or markdown into knowledge bases and get precise, cited answers — in the chat app or over a developer API.
How does docuQ avoid making things up?
Every answer is grounded in your documents and traced to the source chunk it came from. Hybrid dense + sparse (BGE-M3) retrieval with RRF fusion surfaces exact figures — including table numbers that keyword- or vector-only search misses — so answers stay faithful to the source.
What documents can I upload?
PDFs and markdown today. docuQ chunks, embeds, and indexes them automatically; large files are processed in the background so you can keep working while indexing finishes.
Is my data private?
Knowledge bases are private by default and isolated per account — every query is scoped to what you own or have been shared. You can share a KB read-only with teammates, but nothing is public unless you choose to make it so.
Can I use my own model keys?
Yes. Bring-your-own-key (BYOK) lets you point docuQ at your own DeepSeek, OpenAI, OpenRouter, or Anthropic account, with an optional fall back to the house model.
Is there a free trial and a developer API?
Both. Start on a free trial, then move to a paid plan when you are ready. The developer API lets you query your knowledge bases over HTTP with a dq_live_ key and embed docuQ inside your own product.
Start getting answers.
Turn your documents into a precise, cited question-answering engine.