Personal data masked, with a map back.
For teams that send customer or staff text to an AI model.
Looks for personal data in what you send and masks it before it reaches the model. The answer carries a map back to the originals; putting them back is your step.
- 01Your text
- 02Find
- 03Mask
- 04Map back
The answer carries a map back to the originals; putting them back is your step.
Write to [PERSON_1] at [EMAIL_1] about the Q3 report.
Where it sits
The privacy filter looks for personal data in what users send and masks it before it reaches the model. The answer carries a map back to the originals; putting them back in the reply is your step.
Write to Anna Schmidt at anna.schmidt@example.com …
Write to [PERSON_1] at [EMAIL_1] about the Q3 report.
Gets [PERSON_1] and [EMAIL_1] in their place.
You choose what it masks
Each company chooses what it masks.
All personal data
The default.
All but names
All but the kinds you choose
Off
Nothing is masked.
Unless masking is off, it masks the passwords, access keys and card security codes it recognises.
Results on public benchmarks
Share of personal values found, per benchmark.
Loose mode looks for names and usernames, email addresses, phone numbers, postal addresses, dates of birth, personal links and IP addresses, bank and card numbers, ID, passport and VAT numbers, passwords and keys.
Strict mode looks for everything loose mode does, plus any date, codes, places, organisations and job titles.
| Benchmark | Languages | Texts scored | Our privacy filter (strict mode) | Our privacy filter (loose mode) | Open model openai/privacy-filter (default settings) |
|---|---|---|---|---|---|
| REDACT (terms) | 12 EU | 480 | 99.3% | 99.1% | 89.0% |
| SPY | 1 | 1,000 | 99.6% | 98.7% | 86.8% |
| PII-TRACE* | 1 | 250 | 100% | 88.9% | 93.0% |
| Kiji* | 6 | 1,002 | 99.1% | 92.5% | 85.3% |
| Gretel finance* | 7 | 1,001 | 96.3% | 94.0% | 79.1% |
| Gretel English | 1 | 1,000 | 99.8% | 98.8% | 93.7% |
Our filter also masks more text around personal data than the open model: about 2 to 3.5 times the personal data itself, against about 1 to 2 times.
* We used other samples from this benchmark while building our filter. The samples scored here were not used.
REDACT: records from its evaluation split, EU languages only.
SPY: English legal and medical questions.
PII-TRACE (Perplexity): English multi-turn chat.
Kiji (Dataiku): customer support messages.
Gretel finance (Gretel): financial documents.
Gretel English (Gretel): documents from many domains.
The API
How to call it
- 01A person at your company signs in to the console and creates a key. The key is shown once.
- 02POST https://api.astroware.ai/v1/chat/completions, with the header Authorization: Bearer <key>.
- 03Set "model" to "privacy".
- 04The answer carries the masked conversation, a map back to the originals and the kinds found. Send masked_messages on instead. Putting the originals back into your model’s reply with map is your step.
Read the EU Mode API docs before the first call.
{"model": "privacy", "messages": [
{"role": "user", "content": "Write to Anna Schmidt at anna.schmidt@example.com about the Q3 report."}]}{"check": "privacy", "decision": "mask", "personal_data": ["email", "person"],
"masked_messages": [{"role": "user", "content": "Write to [PERSON_1] at [EMAIL_1] about the Q3 report."}],
"map": {"[PERSON_1]": "Anna Schmidt", "[EMAIL_1]": "anna.schmidt@example.com"},
"privacy_mode": "all"}Pricing
- Output tokens are free on all models.
- The same endpoint, key and price as HaloGuard EU.
- Tokens of your text are counted each time they go through a model: text sent to two models, or through two checks in eu-mode, is counted twice.
Start with a $5 trial budget
Start free- For high volumes, a per-token rate agreed for your usage, with invoicing and payment terms to suit.
- Prices can be set in euros.
See what it finds in your own text.
Paste any text. See the personal data our privacy filter finds in it, and the same text with that data masked.
Free. No sign-up. Tested in the 24 official EU languages.
Questions your security team will ask
The privacy filter is part of HaloGuard EU.
The HaloGuard EU page →