Developers, APIs, and automation
Treat site quality like any other release artifact: trigger audits from deploy hooks, store JSON results, and notify teams when scores regress.
Why engineers adopt audit APIs
Marketing asks for Lighthouse scores; security wants headers; leadership wants PDFs. An API lets you centralize those checks and feed the same data to Grafana, Slack, or your admin app.
Give an AI assistant the audit data directly
The workspace speaks MCP, the protocol Claude and other assistants use to call tools, at /mcp. Point a client at it and the assistant gains read tools for sites, audits, reports, per-page results, prioritised opportunities and brand intelligence, scoped to the workspace the credential belongs to.
That means follow-up questions against live data rather than re-exporting a file: which client site regressed this week, which page is dragging a score down, what to fix first. It connects from claude.ai with no key at all, or from Claude Code and Cursor with an API key. Write access is opt-in at approval, uses the monthly audit allowance and refuses rather than spending prepaid credits.
For a fixed snapshot instead, GET /api/audits/export returns many audits as one Markdown, CSV, JSON or NDJSON document. Both need a paid plan; the full walkthrough is in Connect Claude and ChatGPT to your audits.
Where to start
Read the API documentation for authentication, endpoints, and examples. Pair API access with the technical performance solution narrative for what each run measures.
See it in the product
Run an audit on your site to get scores, findings, and AI recommendations in one report.