Tool
Log and code redaction tool
Paste logs or code, flag API keys, tokens, passwords, emails, phone numbers and IPs, then replace them with consistent placeholders — entirely in your browser.
Input
Processing: detection and replacement run entirely in this browser tab — your text is never sent, stored or written into links. Closing the tab discards it.
Rule-based detection can miss or misjudge items — review before sharing.
Review findings
No suspected sensitive items were found — this does not prove the text is safe, please review it manually before sharing.
Redacted output
What gets flagged
The checker looks for known credential formats (OpenAI/Anthropic sk-, GitHub ghp_/github_pat_, AWS AKIA/ASIA, Google AIza/ya29., Slack xox*, SendGrid SG., npm, HuggingFace, Shopify, DigitalOcean, Discord and JWTs starting eyJ), private-key blocks, Authorization / Bearer credentials, key=value style assignments to fields like api_key, token, secret or password (including env-style prefixes such as db_password), passwords embedded in connection URLs, email addresses, phone numbers (+ international format and 11-digit Chinese mobiles) and valid IPv4 addresses.
Field context matters: a bare timestamp or a long random ID is not flagged on its own. The same value under the same category always gets the same placeholder — every occurrence of one email becomes [EMAIL_1]. Overlapping hits are resolved once, quotes and whitespace stay untouched, and manual entries let you cover anything the rules miss. Nothing runs outside this tab; rules can still miss or misjudge, so review before sharing.
FAQ
Is my pasted text uploaded anywhere?
No. Detection, selection and replacement all run inside this browser tab — nothing is sent to this site or any third party, nothing is written to storage or URLs, and closing the tab discards everything.
What can the checker miss?
Rules match known credential formats and common field names. Unusual key names, vendor-specific token shapes, secrets split across lines, and personal data like names, ID numbers or addresses are not covered. A 'no findings' result only means no rule matched — review the text yourself before sharing.
Why do repeated values share one placeholder?
The same value under the same category always becomes the same token — every copy of one email turns into [EMAIL_1]. That keeps the redacted text consistent and readable, and different values still get different numbers.
Does unchecking a finding change other placeholders?
No. Placeholder numbers are fixed when the check runs, so deselecting a false positive keeps every other tag stable — it simply stays in the output as plain text.