Your BotWitness workspace

API key ? Your plan and its access key. The key is like a password that lets you (or your IT team) pull BotWitness records automatically. You don't need it to use this dashboard. Click a revealed key to copy it.

New to the API? See the developer documentation.

Watched domains & alerts ? A watch-list of websites. The moment one of them changes the instructions it gives AI crawlers, BotWitness notifies you, by email, or by posting to a web address if you run automated systems. A single * means "every site we track".

Certificate generator ? Creates a signed PDF exhibit stating what a website told a specific AI crawler on the date you choose, suitable for a filing, an audit, or a negotiation. "Preview verdict" shows the answer free of charge; downloading the certificate uses one credit.

History explorer ? A timeline of every recorded change for a website: what it permitted each AI crawler to do, and exactly when that changed. Think of it as a title history, but for a site's AI-crawler permissions.

Violation check & compliance audit ? Paste web-server log lines (or times and paths) for a bot's visits, and BotWitness checks each one against what the site's robots.txt said at that exact moment, from our timestamped archive. Your log lines are checked and returned, never stored. The audit PDF is the same check as a branded, citable report.

Compliance attestation ? The AI-operator artifact: what robots.txt permitted a named crawler across your domains over a date range, per-capture verdicts from the timestamped archive, as a filing-ready PDF.

GPAI disclosure report ? The EU AI Act Article 53(1)(d) disclosure record: what each watched provider published about its training content, every dated revision, cadence vs the six-month yardstick, and optionally whether your terms appear in the latest disclosure.

Training-data exposure report ? Paste your own text; we probe AI models for verbatim reproduction of it and a calibrated membership signal. An investigative signal, not proof, and never a made-up probability. Uses one report credit (or is included on Monitoring).

Exposure monitoring ? Enroll a work and we re-probe the models when the roster changes (a new flagship ships) or monthly, and email you the day results change. Pair with a copyright canary for the prospective, sound half.

Copyright traps (canaries) ? We publish a unique random canary page and RFC-3161 timestamp it. Because the text is random and its publication date is sealed, a model that later reproduces it demonstrably trained on a post-publication crawl, the one method with a defensible false-positive rate. Prospective: it pays off when future models ship.

Opt-out efficacy report ? Joins the timestamped record of when your domain blocked AI crawlers with live probes of models released since, one PDF: you opted out on date D, and here is what newer models still reproduce or recognize of your text.

Concierge: approval queue ? Inbound email inquiries (litigation packages, defensive audits, enterprise) handled by the AI concierge. Informational replies go out automatically; anything that commits money, evidence, access or a price waits here until you approve it. Nothing below is sent until you click Approve.

Inquiries

Billing ? Your subscription, the card on file, and your receipts, managed right here. Card details are held by Stripe, our payment processor; they never touch BotWitness servers.