Your Codebase, Explained to AI Agents
OpenWiki auto-writes the internal documentation that AI coding agents need to navigate your project — and keeps it fresh every day.
The problem nobody talks about
AI coding assistants — like Claude Code or OpenAI's Codex — are genuinely useful. But they work best when they understand your project's structure. Where things live. What the rules are. How the pieces connect.
Right now, most teams handle this by writing a long instruction file and hoping the AI reads it carefully. It's a bit like handing someone a wall of sticky notes and asking them to build a house.
OpenWiki takes a different approach. It reads your entire codebase and automatically writes a structured wiki — a clear, organised set of notes that explains the project to any AI agent that comes along. Then it connects that wiki to a special file your agents already know to look at.
Here's the part that earns its keep: it also sets up a daily automatic update. Every morning, it checks whether anything in your code has changed and quietly opens a suggested edit if it has. Your documentation doesn't go stale.
For anyone running a dev team — whether it's two people or twenty — this removes a quiet tax that nobody admits is slowing things down.
Words worth knowing
AI coding agent — A piece of software that can read, write, and edit code on its own, following instructions you give it.
Codebase — The full collection of files and folders that make up a software project.
CLAUDE.md / AGENTS.md — Special instruction files that AI coding assistants read first, like a briefing note before they start work.
GitHub Action — A small automated task that runs in the background whenever something happens in your project, like a timer or a trigger.
If you're already using an AI coding assistant with your team, ask whoever manages your code whether you have an AGENTS.md file. If the answer is no — or a blank stare — this might be worth a conversation.