# PII Detection & Redaction

Detect and redact PII in text across European languages: names, emails, phones, addresses, IBANs and more, with positions and scores.

Send up to 100,000 characters of text. You get the text back with personal data replaced by labels such as [EMAIL] or by asterisks, and a list of every span found with its kind, position and score. Works across European languages. Use it before sending text to a language model, a log or a support tool. The model runs on our own hardware in the EU; your text is processed in memory, never stored and never sent to anyone else. A model can miss things: treat this as a strong filter, not a guarantee.

- Finds names and addresses, not only patterns such as emails and card numbers, and works across European languages.
- Text is processed in memory on our own hardware in the EU, never stored and never sent to anyone else.
- Every span comes back with its kind, position and score; up to 100,000 characters in one call.

## Calls

### `POST https://readymadeapi.com/v1/pii/redact`

Send text and get it back with personal data masked, plus where each piece was found and how sure the model is.

- Price: $0.001 per 1k characters, up to 100 per call
- MCP tool: `pii_redact_redact`
- Answers in the same call.

Request:

```json
{
  "text": "Hi, I'm Maria Popescu. Write to maria.popescu@example.ro or call +40 721 234 567."
}
```

Answer:

```json
{
  "redacted": "Hi, I'm [PERSON]. Write to [EMAIL] or call [PHONE_NUMBER].",
  "spans": [
    {
      "label": "person",
      "start": 8,
      "end": 21,
      "score": 0.9513,
      "text": "Maria Popescu"
    },
    {
      "label": "email",
      "start": 32,
      "end": 56,
      "score": 1.0,
      "text": "maria.popescu@example.ro"
    },
    {
      "label": "phone_number",
      "start": 65,
      "end": 80,
      "score": 1.0,
      "text": "+40 721 234 567"
    }
  ],
  "counts": {
    "email": 1,
    "person": 1,
    "phone_number": 1
  }
}
```

## Limits

- Up to 100,000 characters per call.

## Questions

**Which kinds of personal data?**

Person, email, phone number, address, date, date of birth, organization, IBAN, credit card number, bank account number, passport number, ID number, tax ID, IP address, URL, username and password. You can ask for a subset.

**What is it for?**

Anonymization and data masking before text goes to a language model, a log or a support tool.

**Is it a guarantee for GDPR?**

No. A model can miss things. Treat it as a strong filter, not a guarantee.

**How is a call counted?**

Per 1,000 characters, rounded up. A chat message is one unit.

## Using it

- Send `Authorization: Bearer <key>` and a JSON body. Keys and credit: https://readymadeapi.com/account
- Every answer carries `X-Request-Id`, `X-Units` and `X-Charged-Usd`.
- An error is a JSON problem with a `title` to act on and a `detail` in plain words; nothing is charged for it.
- Full schema: https://readymadeapi.com/openapi.json
