What is Lorex?
You use AI coding agents. Maybe more than one. And you've probably noticed that every time you start a session, you end up explaining the same things again — how your project is structured, what patterns to follow, what not to touch.
Then you switch agents, or a teammate joins, and you explain it all over again.
Lorex fixes that.
The idea
Most AI agents already let you give them persistent context — Claude has skills, Cursor has rules, Copilot has instructions. The concept isn't new. The problem is that each tool stores it differently, in a different place, perhaps in a different format.
Lorex doesn't reinvent that. It gives you one place to author that context (a Markdown file called a skill), and then projects it into every agent's native format automatically.
You write it once. Every tool picks it up.
You don't have to start from scratch
Skills can be installed from a team registry or any public Git repository. If your organization already has a registry, one command gets you everything your teammates use. If not, community tap sources let you pull in skills for popular frameworks, tools, or workflows that others have already written.
For solo developers
You write a skill (or ask your AI to write it for you). From that point on, every AI session in this project starts with that context already loaded. No more copy-pasting your conventions into every chat.
For teams
Connect a Git repository as your team's skill registry. Publish a skill once and every developer on the team can install it with lorex install. When the skill evolves in the registry, each developer runs lorex sync to pull the latest. Skills go through the same Git review process as your code, PRs, diffs, history.
For open-source projects
Commit your skills to the repo. Any contributor who has Lorex installed just runs lorex refresh after cloning and their AI agent immediately understands your project's conventions — no setup, no explanation needed.
That's the core of it. Ready to try it?