Docsy

Software / Documentation · Agentic AI

An agent pipeline that reads the codebase, explores the live app read-only, records evidence for everything it sees and checks each claim against both. Only then does it write the manual, every statement linked to its evidence.

[ fig. 1 ]A walkthrough of Docsy: it reads a codebase, explores the live application, records evidence, checks each claim against code and runtime, flags where they disagree, and writes a manual whose statements link to their evidence. Recording, 0:04–0:19.

The problem

Client manuals were written from memory — days of clicking through the product and taking screenshots — and went out of date with the next release. AI writers only guessed faster: fields that don't exist, timings nobody measured.

Before: I think this is how it works
After: I saw this