Documentation
Saer documentation
Set up once, then return whenever your interface changes. Learn how to scan, investigate findings, work with your coding agent and tell us what needs improving.
Back for another scan?
~/.local/bin/saer scan --fullRun this in your project folder. New here? Start with installation and sign-in. Returning? Follow the everyday workflow. Need help? Troubleshoot or give feedback.
Documentation
- Getting startedRequirements, install, first scan, the report.
- What Saer checksEvery capability and its honest state.
- Connect your workflow to cloud historyConnect once, then let normal full scans and CI runs update project history automatically.
- Using Saer day to dayReturn to a project, run a full scan, investigate findings, work with an agent and check changes again.
- Check design-system drift in CICompare committed changes against a base commit, understand exit codes and run Saer on a trusted Mac runner.
- Troubleshooting and feedbackFix sign-in, installation and scan problems, understand free versus full output and send useful beta feedback.
- Reading the reportWhat each part of a Saer report means: exact and inferred findings, unknowns, human decisions and system.json.
- How Saer handles your codeSaer runs locally and read-only. Your code and reports stay local. Optional cloud history saves only the numerical summaries you review. No AI model runs at scan time.
- Saer and AI coding agentsHow Saer gives coding agents your real design system: `saer mcp`, a local MCP server, plus system.json and a check after they write.