Skip to main content

Guides

Treq is an AI code review manager for developers who use coding agents. It gives each task an isolated workspace, keeps those workspaces rebased as the underlying code changes, and lets you review agent output before it reaches your main branch.

Use Treq to inspect diffs like a GitHub PR, leave inline comments, and send those comments back to an agent for revision. Feedback stays tied to the specific code so agents can act on it in the background.

Workspaces are isolated copies of your codebase. Multiple agents can run in parallel without stepping on each other. Treq automatically rebases dependent workspaces, can hand conflict resolution to AI, and supports stacked workspaces for breaking large features into smaller reviewable branches.

Start with Installation to select a repository and create your first workspace. For conceptual background on AI-assisted development and Git workflows, see the Learn section.