We're building the easiest way to read with AI
Documents shouldn't be a wall of text you stare at for hours. They should answer your questions, one conversation at a time.
Knowledge is locked inside PDFs. We're handing you the key.
— The DocChat team
By the numbers
0M+
Documents processed
0K
Active users
0+
Countries
0+
Languages
From a frustrated Friday night to 85,000 users
- 1
2024
DocChat is born
Three engineers, one frustrating Friday night reading a 200-page paper, one idea: "what if you could just ask?"
- 2
2025
Citations as first-class
Built our retrieval pipeline from scratch so every answer ships with the exact page and section it came from.
- 3
2025
Hit 50,000 users
Word-of-mouth from researchers, lawyers, and analysts. Zero ads. The product was the marketing.
- 4
2026
Multi-document reasoning
Released cross-document chat — ask one question, get answers synthesised across a whole library.
Four principles we won't compromise on
Accuracy first
Every answer is grounded in the document. If we don't know, we say so — no hallucinations, no guessing.
Built for humans
Design that disappears. We obsess over the moment between you having a question and getting an answer.
Open by default
Public roadmap. Transparent pricing. Export your data anytime. No vendor lock-in, ever.
Team-friendly
From solo researchers to enterprises — collaboration features are the same on every plan.
A small team that ships a lot
We're researchers, engineers, and designers who've all been on the wrong side of a 300-page PDF at 11pm. We built DocChat for ourselves first.
Researchers
Half our team has a research background. We build for the bench.
Shippers
We ship weekly. Big ideas matter. Boring details matter more.
Privacy hawks
Your documents are yours. We don't train models on them. Ever.
Ready to read with superpowers?
Upload your first document in under a minute. Free forever — no credit card, no trial gimmicks.