OCR for PDFs, images, Word and PowerPoint

Key facts

Input

PDF, images (PNG, JPEG, WebP, AVIF), DOCX and PPTX – as upload or URL

Output

Markdown or plain text; embedded images as download links, base64 or left out

Costs

6 credits per page; the balance is checked before the document is processed

Pages

Up to 8 pages per call, selected with the pages parameter

Processing

In the EU by Mistral AI, no training on your data

Available via

REST API (https://api.1001fx.com) with your API key – from Make, n8n, scripts and AI agents

What you can do with it

• Make scanned contracts, letters and forms searchable
• Feed documents as Markdown into an AI step, a knowledge base or a summary
• Read the text of photos, screenshots and whiteboards
• Turn Word and PowerPoint files into text for translation or review

How it works

1. 1

   Pick the function for your file type: ocrpdf, ocrimage, ocrdocx or ocrpptx.

2. 2

   Send the file as upload or as fileUrl, optionally with the pages you need.

3. 3

   Choose Markdown or plain text as output.

4. 4

   Use the text in the next step of your workflow.

Functions

Each function has its own page with all parameters, examples and a ready-to-copy curl command.

Example

A call of https://api.1001fx.com/files/ocrpdf. Replace YOURSECRETAPIKEY with your API key.

curl -X POST 'https://api.1001fx.com/files/ocrpdf' \
  -H 'X-API-KEY: YOURSECRETAPIKEY'

Frequently asked questions

Does OCR work with scanned PDFs and handwriting?

Yes. The functions read scans, photos, tables and handwriting. The quality depends on the resolution of the scan.

How do I process documents with more than 8 pages?

Call the function several times with the pages parameter, for example 1-8, then 9-16.

Where are my documents processed?

In the EU by Mistral AI. Mistral does not train on the documents. 1001fx does not store uploaded files in the call log.