AI prompt privacy

Anonymise text before using ChatGPT or AI tools.

SanitiseAI helps turn sensitive notes, emails, transcripts, and document excerpts into safer AI prompts by replacing personal and operational identifiers with readable placeholders.

What to check for

  • Names and personal contacts
  • Addresses and dates of birth
  • Customer IDs and invoice references
  • Usernames and account identifiers

What to preserve during review

  • Core facts
  • Structure and chronology
  • Task-relevant context
  • Readable placeholder labels

In practice

A worked example

Synthetic input and expected readable output. This small example demonstrates one replacement, not complete coverage. Try it in the tool and review the result.

01 / Source text

Please rewrite this escalation. Email alex@example.com about the delayed delivery; the customer needs an update today.

02 / Expected output

Please rewrite this escalation. Email [Email 1] about the delayed delivery; the customer needs an update today.

Follow the workflow

  1. Remove attachments and unrelated conversation history before pasting.
  2. Run automatic detection, then check the output against the source.
  3. Paste only the reviewed output into your assistant, followed by your rewriting instruction.

Put the workflow into practice.

Start with an example. Review every result before sharing.

Open sanitiser