Protect prompts before AI
Mask names, emails, phone numbers, addresses, IDs, secrets and other sensitive values before copying text into an LLM.
DATA PROTECTION LAYER FOR LLM WORKFLOWS
Online text anonymizer for AI prompts, LLM data masking and reversible text protection.
Protect sensitive text before you paste it into AI.
anonymise.online is a reversible data-hiding layer for LLM use. It detects personal data, credentials, business identifiers and confidential text, replaces them with readable placeholders, lets you review the result, and exports a private restore map that you keep under your control.
Free: 500 characters · 2 anonymizations/day · no signup required.
Readable placeholders
Keep context with labels like PERSON, EMAIL, ADDRESS or TOKEN so the prompt still makes sense.
Manual phrase control
Select a full phrase directly in the preview or add exact terms. Selected phrases become dedicated MANUALBLOCK placeholders.
Restore after AI
When you get the model output back, use your private restore map to put the original values back into place.
TRUSTED FLOW
Protect → use AI → restore.
Use anonymise.online as the layer between raw text and your preferred AI tool. It is not a VPN, browser proxy or AI gateway: it simply protects text first, then lets you decide where the anonymized version goes.
See the full workflow