Protect sensitive data before sending it to an LLM
I built Anon to remove PII from text and PDFs before they are sent to tools like ChatGPT or Claude.
It detects sensitive data automatically, lets you review or manually protect anything it misses, and replaces it with reversible placeholders such as:
[EMAIL_A12F_REDACTED]
[PERSON_B83C_REDACTED]
[MANUALBLOCK_D91A_REDACTED]
You keep a private JSON map so the original values can be restored later.
It supports both text and PDFs, with no account required for basic use.
I’m looking for feedback, especially on detection accuracy and the PDF workflow.
Top comments (0)