Does it support two-way restoration?

Yes, temporary local mappings allow restoring original secrets once the AI generates an answer.

Which cloud tokens are recognized?

OpenAI, Anthropic, GitHub, AWS, Google Cloud, Stripe, and Slack API keys.

MiniToolDesk - 100% Free & Private Online Developer Tools

100% In-Browser Privacy Protection

All sanitization occurs 100% client-side via JavaScript. Zero network requests, zero data logging.

AIR-GAPPED COMPLIANT

Active Redaction Rules (Click to toggle)8 Hazards Redacted

Original Unsafe Code / Logs / JSON577 chars

Sanitized Output (Safe to Send to AI)

[Notice: Sensitive data (keys, tokens, IPs, credentials) has been sanitized as <TAG_N> for privacy compliance. Please keep all <TAG_N> placeholders intact in your analysis/code response.]

[ERROR] 2026-10-14 09:23:45 UTC
Database connection failed on <DB_URL_1>
Caller IP: <IP_ADDR_1> (Gateway: <IP_ADDR_2>)
User ID: 884920, Name: Alice Wang, Phone: <PHONE_1>, Email: <EMAIL_1>
Failed to authorize OpenAI request with key: <API_KEY_1>
Session JWT: <JWT_TOKEN_1>
AWS Fallback Key: <API_KEY_2>

How to sanitize prompts before ChatGPT/Claude

1Paste code, stack trace, or JSON logs into the input editor.
2Review detected risk categories: API Keys, JWTs, IPs, and PII are redacted instantly.
3Click "Copy for AI" to feed the safe placeholder text to LLMs.
4Switch to Tab 2 to de-anonymize AI responses back to original values.