Restore original values with the mapping file
Upload an anonymized CSV (or paste text from an AI assistant) plus the mapping dictionary. The tool replaces every token with its original value. The reverse never touches our servers; the mapping is your key.
When to use de-anonymize
You anonymized a dataset to share with an external team or to send to ChatGPT for analysis. Now you have the analysis back. But the rows reference tokens like EMAIL_0001 or NAME_0023 that you need to translate back to the real customer records.
Reverse anonymization is also useful when you no longer need the dataset to be anonymized: you can restore it for archival or merge it back into a production workflow as long as you kept the mapping dictionary.
How it works
- 1Upload anonymized fileCSV, TSV or Excel produced by the anonymize tool. Or switch to the 'pasted text' tab to reverse LLM output.
- 2Upload mapping dictionaryThe JSON or CSV file the anonymize tool downloaded alongside the result.
- 3Reverse and downloadThe tool replaces every mapped token with the original value and gives you the restored file or text.
Need to anonymize a fresh dataset?
Use the anonymize tool first to produce both the anonymized file and the mapping dictionary. The mapping is the key. Keep it safe.
Open the anonymize tool →Frequently asked questions
Do I need the original anonymized file?+
For file mode, yes. The tool replaces values in the file using the mapping. For pasted text mode, you only need the mapping plus the text you want to reverse (e.g. an LLM response).
What if the mapping file is incomplete?+
Tokens not in the mapping are kept as-is. Partial reversal is fine. The rest of the file is unaffected.
Is the mapping file uploaded anywhere?+
No. The mapping is parsed in your browser, used locally to replace tokens, and never sent to our servers. The whole reverse is offline-capable.
How much does it cost?+
Reverse anonymization is free, with no row limits. Anonymization is what costs above 2,000 rows; reverse is always free.