Getting started

1. Preview audit

Enter your URL on the audit page. Up to 10 public HTML pages are crawled with robots.txt respected. The preview includes Search Health, the top findings and a detected profile.

2. Create a workspace

Signing up creates a workspace and a project. The free plan is enough to start: 1 project, 10 pages per crawl.

3. Full audit and plan

Press Run audit. The engine crawls, stores findings with evidence, scores them, clusters keywords and generates AI monitoring prompts.

4. Execute and verify

Fix findings, produce content, then re-crawl. A fixed finding must disappear in the next crawl; that is the only real measure of success.

What each run actually measures

Each crawl reads public HTML pages and records the server response, index directives, title and description, heading outline, internal and external links, images, structured data and hreflang. Those same observations are then scored against a published rubric. Anything not measured stays labelled "not measured"; ranking, traffic and AI answers are never claimed without real data.

After a run, check three things: does each finding carry concrete evidence, is the recommended action executable, and does the finding disappear on the next crawl once fixed. If any of the three fails, treat the report as a list of questions rather than a decision.

Implementation status

This table reflects actual state, not intentions:

Local setup

Requires Node.js 20+ and SQLite (bundled with the OS).

npm install
cp .env.example .env      # generate AUTH_SECRET and ENCRYPTION_KEY
npx prisma migrate deploy
npm run dev               # web on :3000
npm run worker            # background jobs (optional second terminal)

For PostgreSQL, see DATABASE.md in the repository root.