Paste redact for AI

Client

Replace common sensitive patterns with [REDACTED] so you can share context with an LLM or teammate more safely. For JSON documents, use JSON redact keys when you need to mask by field name.

Heuristic patterns

Toggle categories to control what is stripped: email addresses, phone-like numbers, JSON Web Tokens (three dot-separated segments), Bearer authorization lines, URLs that embed user:password@, and a few vendor prefixes (for example OpenAI sk-, Google AIza…, AWS AKIA…, GitHub and Slack tokens). The tool reports how many matches it replaced—double-check the result every time.

Plain text

?

Patterns are heuristic (regex). Review the output—secrets can hide in unexpected formats, comments, or base64. For JSON payloads, also use JSON redact keys.

Redact

Replacements this pass: 5(email 1, phone 1, jwt 1, bearer 0, URL 1, vendor 1)

Common use cases

  • Scrub a log excerpt or ticket body before pasting into a chat assistant or public forum.
  • Strip obvious Bearer and sk- lines from curl copies while keeping structure readable.
  • Pair with JSON key redaction when the payload is still structured JSON.

Common mistakes to avoid

  • Assuming every secret was caught

    Custom formats, hex blobs, or renamed env vars can slip through. Read the full output and remove anything that still looks sensitive.

  • Redacting instead of rotating

    If a real credential was exposed, rotate it at the source—redacted text does not revoke access.

FAQ

Is this as good as a DLP appliance?

No. It is a fast, offline helper with common patterns. Use enterprise tools when policy requires guarantees.

Does text leave my device?

No. Matching runs entirely in your browser.

More tools

Related utilities you can open in another tab—mostly client-side.