Prompt & paste sanitiser
Redact sensitive data before ChatGPT and other AI assistants: paste a draft below to remove API keys, emails, and IBAN-like patterns from your text. Highlights show what was matched; the redacted box is safe to copy into an AI chat. Everything runs locally in your browser — we do not upload your text.
Why redact sensitive data before using AI?
Large language models may log or train on prompts depending on the product and your settings. Sending API keys, bank details, or personal identifiers in a chat is risky. This page helps you remove API keys from text before AI tools see it and spot common patterns so you can paste a cleaner prompt. It is a best-effort helper — always review the redacted output and treat any exposed secret as compromised.
What we try to detect
The sanitiser looks for email addresses, IBAN-style bank identifiers, PEM private key blocks, JWTs, and several common token shapes (for example OpenAI sk- keys, AWS-style access keys, GitHub and Google API tokens, Stripe keys, Slack tokens, and Telegram bot tokens). Unknown or custom secrets may still slip through — use judgement and your organisation’s policies.
Frequently asked questions
- How do I redact sensitive data before ChatGPT?
- Paste your text here, review the highlights, then copy the redacted version from the output box into ChatGPT or another assistant. Nothing is sent to our servers.
- How do I remove an API key from text before sending it to an AI?
- Paste the text; known key shapes are replaced with
[REDACTED:…]placeholders. If a key was ever pasted into an untrusted place, rotate it in the provider’s dashboard. - Does this tool upload my text?
- No. Processing runs in your browser only.