Add document anonymization to your own workflow with a REST API. Upload a PDF, Word file, text file or scan, and Re-Doc finds the personal and health information in it, then redacts it or replaces it with realistic synthetic data while keeping the original layout. Three calls: upload, poll, download.
Synthetic replacement for text-searchable PDF, DOCX and TXT: names, dates, IDs, addresses and contact details are replaced with consistent synthetic values, and the output keeps the same format and layout.
Redaction for scanned PDFs, images (PNG, JPG, TIFF, WebP, BMP, GIF) and DOCX: sensitive text is read with OCR and permanently blacked out.
Multiple data types (Pro and higher) for a DOCX with text and images: text is replaced and logos, stamps and signatures inside images are redacted in one pass.
Personal and health identifiers that appear as text: names, dates, addresses, phone numbers, emails, ID and account numbers and medical record numbers. Add your own entity types with an optional custom prompt.
One file per request, up to 50 MB (5 MB for TXT). Send requests in parallel from your own job queue; each upload returns a session ID to poll. Check page usage with GET /api/v1/usage/check.
All calls use HTTPS with bearer API keys. Documents and outputs are typically kept up to 30 days and then deleted, and are not used to train Re-Doc's own models (see the privacy policy). On-premise deployment is available on request, so documents stay on your infrastructure.
The API is included on paid plans. Try the same processing free in the web app (first 10 pages).
Authorization: Bearer re-doc-xxx...POST /api/v1/processing/upload file: document.pdf
GET /api/v1/processing/
status/{session_id}GET /api/v1/processing/
download/{session_id}/converted/0Processing thousands of documents? Our REST API handles batch redaction, text replacement, and multiple data types at scale.
Upload a scanned PDF, image or DOCX. We find the sensitive data and black it out.
POST /api/v1/processing/uploadUpload a native PDF, DOCX or TXT. We replace sensitive values with synthetic data that looks real and keep your layout.
POST /api/v1/text-replacement/process-searchable-pdfGet your processed document back. Text replacement keeps your format: PDF stays PDF, DOCX stays DOCX, TXT stays TXT. Redaction returns a PDF, plus a DOCX for DOCX uploads.
GET /api/v1/processing/download/{session_id}/converted/0Send a DOCX that mixes text and images. We replace the sensitive text and redact the sensitive parts of embedded images in one pass, keeping your layout intact.
POST /api/v1/hybrid/uploadYour API key is your access pass. Create one, add it to your requests, and you're ready to process documents programmatically.
Create API keys to access Re-Doc programmatically.
Note: API access requires a paid plan (Starter or higher).
Sign in with Googlecurl -X POST https://re-doc.com/api/v1/processing/upload \ -H "Authorization: Bearer re-doc-your-api-key" \ -F "file=@document.pdf" \ -F "lang=en" \ -F "document_type=general"
curl -X POST https://re-doc.com/api/v1/text-replacement/process-searchable-pdf \ -H "Authorization: Bearer re-doc-your-api-key" \ -F "file=@notes.txt" \ -F "txt_mode=redact" \ -F "txt_redact_style=tag"
Upload a file and we'll figure out the rest. But if you need control, here's what you can tweak.
| Parameter | Required | Description |
|---|---|---|
| file | Yes | The document to process, one file per request, up to 50MB. Redaction: scanned PDF, DOCX, PNG, JPG, TIFF, WebP, BMP, GIF. Text replacement: text-searchable PDF, DOCX, TXT. Multiple data types: DOCX. |
| lang | No | Language code for OCR processing. Defaults to "en" (English). Common values: en, es, fr, de, it, pt, zh, ja, ko |
| document_type | No | Document category to optimize entity detection patterns. Options: "general", "financial", "medical", "legal" |
| custom_prompt | No | Custom instructions to focus entity detection on specific information types. Example: "Focus on social security numbers and dates of birth" |
| txt_mode | No | TXT files only. "replace" (default) swaps sensitive values for realistic synthetic data; "redact" removes them and marks where they were. |
| txt_redact_style | No | Used when txt_mode is "redact". "tag" (default) writes labels such as [PERSON_NAME] or [EMAIL]; "block" writes ████ blocks. |
Our OCR reads every character, even from tilted scans and faded prints. Works across 80+ languages.
The tool finds names, Aadhaar numbers, account details, and whatever else you tell it to look for.
Redaction marks go exactly where they should. Your document comes back in the same format you sent it.
/api/v1/processing/uploadAuth requiredUpload a document for redaction (black-box masking of sensitive data)Content-Type: multipart/form-dataUpload a scanned PDF, an image (PNG, JPG, TIFF, WebP, BMP, GIF) or a DOCX. One file per request, up to 50MB. The API uses OCR to read the text and places black boxes over detected PII. Native PDFs, DOCX and TXT files that should keep their text can use the text replacement endpoint instead.
{
"file": "document.pdf, document.docx, or image.png (File) - Required",
"lang": "en (string) - Language code for OCR (default: \"en\")",
"document_type": "general (string) - Document type: \"general\", \"financial\", \"medical\", \"legal\" (default: \"general\")",
"custom_prompt": "Detect PII in legal documents (string, optional) - Custom prompt for specific entity detection"
}{
"session_id": "91f53136-a734-4fb9-a8b8-482d1de14035",
"status": "pending",
"message": "Document uploaded successfully. Processing started.",
"file_name": "document.pdf",
"estimated_time": 120
}{
"detail": "File type not supported: text/plain"
}{
"detail": "File too large. Maximum size: 50.0MB"
}/api/v1/processing/status/{session_id}Auth requiredPoll the processing status of a redaction job until status is "completed" or "failed"{
"session_id": "91f53136-a734-4fb9-a8b8-482d1de14035",
"status": "completed",
"progress": 100,
"message": "Status: completed",
"error_message": null,
"created_at": "2026-10-01T16:26:13",
"updated_at": "2026-10-01T16:26:42"
}/api/v1/processing/results/{session_id}Auth requiredGet the results of a completed redaction job: counts and the list of output files{
"session_id": "91f53136-a734-4fb9-a8b8-482d1de14035",
"status": "completed",
"file_name": "document.pdf",
"total_images_processed": 1,
"total_entities_found": 12,
"total_pii_entities": 12,
"redacted_images": [
"…/page-1_redacted.png"
],
"converted_documents": [
"…/document_redacted.pdf"
],
"created_at": "2026-10-01T16:26:13",
"completed_at": "2026-10-01T16:26:42"
}{
"detail": "Processing not completed"
}{
"detail": "Processing session not found"
}https://re-doc.comAll requests go here. HTTPS only, no exceptions.
Extra pages on Pro and Business: $0.02 each, after approval (contact us first). Multiple data types: Pro and higher. See pricing.
Need more? Email support@re-doc.com
A service you call from your own code to find personal data in documents and remove or replace it, so files can be shared, analysed or used for AI without exposing people.
PDF, DOCX and TXT for synthetic replacement; scanned PDFs, images (PNG, JPG, TIFF, WebP, BMP, GIF) and DOCX for redaction. Files can be up to 50 MB (5 MB for TXT).
Yes. Synthetic replacement returns the same format with the layout kept. Redaction returns a PDF, plus a DOCX for DOCX uploads.
Each request takes one file. To process many documents, run requests in parallel from your own queue and poll each session.
On-premise deployment is available on request through an enterprise discussion. Contact our team to discuss your requirements.
Integration issues, edge cases, or just want to talk architecture? We're here.