Prompt privacyStrip the secrets
Strip the secrets
before the prompt.
A free, in-browser tool that redacts sensitive data from your prompts. Nothing is uploaded.A free, in-browser tool that redacts sensitive data from your prompts so you can use any AI safely. Nothing is uploaded.
Anonymized
Your safe-to-send prompt appears here. Placeholders like [EMAIL_1] replace your real data; the mapping below lets you restore them in the AI's reply.
Heuristic detection: scan the result before sending. Toggle categories above to tune it.
What it catches
The data you shouldn't paste.
It scans for the categories of data most likely to be confidential or personal, and replaces each with a placeholder you control.
- Identities: names, @handles, and email addresses.
- Contact and records: phone numbers and customer details.
- Company data: organization names and business identifiers.
- Secrets: API keys, tokens, passwords, and Bearer credentials.
- Financial and government IDs: card numbers (Luhn-checked), IBANs, and SSNs.
- Network: IP addresses and URLs.
FAQ
Prompt privacy: common questions.
How the tool protects your data and where its limits are.
Go further
Want AI your data can actually trust?
We design and run AI with privacy built in: data-handling policies, model choice, redaction in the pipeline, regional hosting, and on-prem or self-hosted options when the data demands it.
Free downloads
delivered by email.
Pick a document and we email you the PDF.
Safe Prompting: scrub before you sendWhat each PII category catches, why raw data does not belong in prompts, and a pre-send checklist.