Stop losing the “why”
Review comments scroll away, threads vanish, and the reasoning lives in one person's head — until they leave or an agent's context resets. Doco keeps it.
New contributors inherit the reasoning behind the code — not just the code itself.
Before they act, agents search the Doco — so they build on prior decisions instead of contradicting them.
Turnover, re-orgs, and reset context windows no longer erase your institutional memory.
The same structured record powers your team and every agent working alongside them.
Not another wiki — a living graph
Doco captures structure, relationships, and provenance that prose docs can't.
Decisions, Rules, Intents, Actions, and References — each a first-class node with the right shape, never a wall of prose.
Edges link a decision to the rule it spawned and an action to who performed it — trace cause and effect, don't just read pages.
A native MCP server, OAuth 2.1, and a shared protocol let Claude Code, Cursor, and Codex discover and query your Doco automatically.
Ask in plain language and get the ranked decisions and rules that actually bear on the question.
Business-process maps, glossaries, and org trees — live views built from the same nodes, not diagrams that rot.
Link any node to the exact file and line that implements it, on a history that's only ever added to.
Connect, capture, recall
Run the wizard or doco install-agent-bootstrap. Doco commits a few files so every teammate and agent finds the same memory.
As decisions get made and rules emerge, you (or your agent) just "doco it." The why lands as a typed node the moment it forms.
Anyone — human or agent — searches before acting, and builds on what the project already knows.
Your AI can read this too
Doco is as usable by your agents as by you. Point any agent at /llms.txt and it self-onboards — authenticates, reads the protocol, and starts contributing to the shared record. Zero install.