CLI
cinder integrate
Crawl a real running instance of your app, audit every page for untagged elements and weak labels, and write one consolidated report.
Need the shortest path?
Use Quick Start when you want the smallest working route → scan → graph → run flow first.
Something feels off?
Inspect scan output, graph state, active route/context, and dispatch in a practical order.
Need the full surface?
Browse all runtime methods, hooks, extension points, and lifecycle APIs.
npx cinder integrate <url> [options]Crawls a real running instance of your app (same-origin pages only, starting from <url>), runs both of Cinder's tagging-checklist audits on every page it finds, and writes one consolidated report - meant to be handed to a coding agent (Claude Code, Cursor, or similar) to apply.
Options
| Option | Default | Description |
|---|---|---|
--max-pages <n> | 25 | Stop after this many pages. |
--settle-ms <n> | 400 | Pause after each page load before auditing it. Real testing found rapid sequential navigation can catch a page before its accessible name settles - don't lower this unless you've verified it's still accurate. |
--storage-state <f> | - | Path to a Playwright storage-state JSON file (cookies + localStorage) for an already-authenticated session. |
--out <file> | cinder-integration-report.md | Where to write the report. |
Authenticated apps
Without --storage-state, any page behind a login wall just redirects to your login page and audits that instead of the real app - found by actually running this against a real app, not a guess. Export a storage-state file once after logging in:
await context.storageState({ path: 'storage-state.json' });or via:
npx playwright open <url>Then pass it:
npx cinder integrate https://app.example.com --storage-state storage-state.jsonWhat you get
Crawling https://app.example.com (up to 25 pages)...
[1/12] https://app.example.com/
[2/12] https://app.example.com/reports
...
Scanned 12 page(s): 4 untagged element finding(s), 2 weak-label finding(s).
Report written to cinder-integration-report.md
Hand that file to your coding agent to review and apply.The report is meant to be read by a coding agent, not applied automatically - the same honest boundary every other Cinder audit tool keeps: this reads your app, it never edits it.
Continue reading
Previous: CLI Overview
cinder integrate crawls your real running app and audits every page for untagged elements and weak labels; cinder install-skill teaches your coding agent to keep up.
Next: cinder install-skill
Install a self-contained Claude Code Skill and an AGENTS.md section, so your coding agent runs the tagging checklist without being told each time.