# Anonymize a PDF: replace personal information with labels

> Upload a PDF or scan. Reducto finds names, postal addresses, SSNs and other ID numbers, account, card, routing and IBAN numbers, phone numbers, email addresses, dates of birth and signatures, and replaces each one with a label such as [PERSON 1] or [ADDRESS 2]. The same person or address gets the same label everywhere, so the document still makes sense. The pages keep their layout, the original text under each label is removed, and the file is checked again before you download it.

Canonical: https://reducto.ai/tools/anonymize-pdf  
Last updated: 2026-11-15  
For: Teams sending documents to ChatGPT, Claude or other AI tools, Researchers sharing case files, Support and ops teams sharing customer documents

## Run it now (free, no API key, up to 5 pages)

Send a PDF, Word, PowerPoint, Excel or image (PNG, JPG, HEIC, TIFF, BMP, GIF) (max 10 MB, 5 pages) to `POST https://reducto.ai/api/tools/anonymize-pdf` as multipart `file`, or JSON `{"url": "https://..."}`. The response has `result`, ready-to-save `output` (Anonymized PDF (same layout, labels in place, original text removed), Findings with their labels (JSON), Leak check (checks)), `pages_processed` and `next`.

```bash
curl -s https://reducto.ai/api/tools/anonymize-pdf \
  -F file=@bank-account-letter.pdf

# or, for a file that is already online:
curl -s https://reducto.ai/api/tools/anonymize-pdf \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com/bank-account-letter.pdf"}'
```

Higher volume needs a free API key (new accounts get 15,000 free credits). Agents can sign up over HTTP: https://reducto.ai/auth.md. Humans: https://studio.reducto.ai/?utm_source=free_tool&utm_medium=agent&utm_campaign=anonymize-pdf&utm_content=api_response

Big file? Without a key the free run covers the first 5 pages and returns `partial: true` with a `signup_required` block. Your user signs up in under a minute (email, 6-digit code, name, terms: https://reducto.ai/auth.md); then repeat the same request with `Authorization: Bearer $REDUCTO_API_KEY` for the whole file on their 15,000 free credits. The same key works on the hosted MCP (https://mcp.reducto.ai/mcp) and the Reducto API.

## How to do it

1. Upload a PDF, a scan or a photo of a document.
2. Reducto finds the personal information: names, postal addresses, SSNs and other ID numbers, account, card, routing and IBAN numbers, phone numbers, email addresses, dates of birth and signatures.
3. Each person, address and number gets a label, such as [PERSON 1], and the same one gets the same label everywhere.
4. Review the list and untick anything you want to keep, then each page with a label is rebuilt as an image with the label in place of the original text.
5. The finished PDF is read again to check none of the original values can be found, then you download it.

## How it works

POST the file with action=detect to get the findings (each with its category, text, page and boxes), then POST it again with action=apply and the findings you keep (JSON), plus optional custom phrases and boxes. With no action the tool detects and applies in one call. The output is only returned when the leak check passes: checks[0].status is pass, or download_blocked is true with the leaked values listed.

| `model` | `r-1` | Reads typed text, scans, photos and handwriting, with a box for every word. |
| `return_ocr_data` | `true` | Word-level boxes, so each label goes exactly where the text was. |

### curl

```bash
curl -s -X POST https://reducto.ai/api/tools/anonymize-pdf \
  -F "file=@letter.pdf" -F 'custom=["Project Falcon"]' \
  | tee result.json | jq '.checks'
jq -r .output.pdf_base64 result.json | base64 -d > letter.anonymized.pdf
```

### Python

```python
import base64, requests

res = requests.post(
    "https://reducto.ai/api/tools/anonymize-pdf",
    files={"file": open("letter.pdf", "rb")},
    data={"custom": '["Project Falcon"]'},
)
res.raise_for_status()
body = res.json()
print(body["checks"])
if not body["output"]["download_blocked"]:
    open("letter.anonymized.pdf", "wb").write(base64.b64decode(body["output"]["pdf_base64"]))
```

### Node

```ts
import { openAsBlob } from "node:fs";
import { writeFile } from "node:fs/promises";

const form = new FormData();
form.set("file", await openAsBlob("letter.pdf"), "letter.pdf");
form.set("custom", JSON.stringify(["Project Falcon"]));
const res = await fetch("https://reducto.ai/api/tools/anonymize-pdf", { method: "POST", body: form });
const body = await res.json();
console.log(body.checks);
if (!body.output.download_blocked) await writeFile("letter.anonymized.pdf", Buffer.from(body.output.pdf_base64, "base64"));
```

## What it handles

- **Still readable**: "[PERSON 1] opened account [ACCOUNT 1]" still tells the story, without saying who.
- **The same label everywhere**: Each person, address or number keeps one label across every page.
- **Same layout**: Labels sit where the original text was, and the original text is removed from the file.
- **Scans and photos**: OCR reads every word, so scanned pages are anonymized like typed ones.

## FAQ

### How is this different from redacting?

Redacting blacks the information out. Anonymizing replaces it with a label, such as [PERSON 1] or [ACCOUNT 2], so you can still tell who did what in the document.

### Does the same person get the same label?

Yes. Every mention of the same name, address or number gets the same label across the whole document, including first or last names on their own.

### Can it make up realistic names and addresses instead?

Not this free tool: it only uses labels. If you need realistic replacement data that keeps documents useful for AI training or sharing, see document anonymization at /use-cases/document-anonymization.

### How do I know the text is really gone?

After the file is built, Reducto Parse reads it again, both the text layer and OCR of the page images, and every redacted value is searched for, including its digits without spaces or dashes. You see "Verified: none of the N redacted items can be found in the output". If anything is still readable, the download is blocked and the page lists what leaked.

### Does it work on scans?

Yes. Reducto reads scans, photos and handwriting with OCR, so personal information on a scanned page is found and labelled too.

### Is it free?

Yes. Short documents run free with no account, with a small daily limit per IP. For a longer file you get a free preview of the first pages, then enter your email to get the whole file on a free account (15,000 free credits).

Sample document (synthetic): https://reducto.ai/tools/samples/redact/bank-account-letter.pdf

## Related

- [Redact a Word document](https://reducto.ai/tools/redact-word-document)
- [Redact a PDF](https://reducto.ai/tools/redact-pdf)
- [Make a scanned PDF searchable](https://reducto.ai/tools/ocr-scanned-pdf)
- [PDF to text](https://reducto.ai/tools/pdf-to-text)
- [All Redact & anonymize tools](https://reducto.ai/tools/redact)
