OCR PDF to JSON

Turn scanned PDFs and images into structured JSON. Give the actor a list of document or image URLs and it returns the extracted text, detected tables and key-value pairs, one row per page, with the source url on every row. It runs Tesseract OCR inside the actor, so it needs no API keys and no external services.

Who it is for

• AI agents and LLM pipelines that need text from a scanned file before reasoning over it.
• Developers who want a cheap, pay-per-page OCR step without running their own servers.
• Operations teams digitising invoices, forms, receipts and letters.

What each row contains

Each dataset row is one page. It has the source url, the page number, the extracted text, and any tables and key-value pairs found. The output is flat and ready for a spreadsheet, a database or a prompt. Open the Output tab after a run to see the exact columns.

Three use cases

1. Invoice and receipt intake: extract vendor, date and totals from scans, then send them to accounting.
2. Agent document reading: an AI agent calls the actor through MCP to read a scanned PDF it cannot parse otherwise.
3. Archive digitisation: batch old scanned letters and forms into searchable JSON.

How to use

1. Open the actor and paste one or more PDF or image URLs into the input.
2. Click Start. The default input runs a small sample.
3. Download the results from the dataset as JSON, CSV or Excel.

API (curl)

curl -X POST "https://api.apify.com/v2/acts/hamsmoothie~document-ocr-to-json/run-sync-get-dataset-items?token=YOUR_TOKEN"
-H "Content-Type: application/json"
-d '{}'

An empty body uses the default input. Add your own URLs using the field names shown in the Input tab.

MCP (AI agents)

The actor is callable from Apify's MCP server. Search for "ocr pdf to json" and call hamsmoothie/document-ocr-to-json with your document URLs.

Pricing

Pay per result: $2.00 per 1,000 pages ($0.002 per page). You only pay for rows returned.

FAQ

Which formats work? Common image formats and PDFs reachable by a public URL.

Is my data stored? Results live in your own Apify dataset. The actor only fetches the URLs you supply.

How accurate is it? Tesseract works well on clean printed text. Handwriting and very low-quality scans give weaker results.

Can it read password-protected files? No. Only public, unprotected URLs.

Related compliance tools

Other actors from Maybe Quark: Contractor License Verifier, FMCSA Carrier Safety Lookup, Secretary of State Business Entity Search, Healthcare Professional License Verifier, NPI Registry Lookup, Exclusion & Debarment Screener, FDA Recalls, DOL Wage & Hour Violations.