Trust

Privacy

Short version: we keep as little as possible, for as short as possible.

What we store and for how long

Your email, so we can send the result and keep track of daily limits and credits.
The input you send (URL, file or text) and the result. Both are deleted automatically after 7 days.
A short log of the steps the agent took, shown on the job page. Deleted with the job.
The sandbox the agent ran in is destroyed immediately after the job.
Google Analytics counts page visits (pages viewed, country, device and browser) so we can see which agents people use. It runs and sets cookies only if you click "Allow"; you can change this any time under "Cookie settings" at the bottom of the page. We never send your email, check results or the private part of report links to it.

Agents use the Anthropic API (Claude) to explain and summarise. What is sent: the script text, page text and metadata of audited pages, or the API specification you provided.
Anthropic does not use API data to train models. We never send your email to the model.
Content from websites and scripts is treated strictly as data; the agent has no tools that can act outside its sandbox.

Only audit websites you own or are allowed to check. The auditor is rate limited, respects robots.txt and cannot reach private networks.
Do not upload real API keys or secrets. Generated MCP servers read keys from environment variables.
Results are produced by automated agents and may contain mistakes. A clean script report is strong evidence, not a guarantee.

Payments are processed by Lemon Squeezy, which acts as merchant of record and handles VAT. We never see your card details.
If a job fails, the credits it used are returned automatically.
Refunds are described in the terms.