Introduction
dOCR turns documents into structured data and renders web pages into images — one API, one set of credits.
dOCR is two products on one API. Extract turns documents into structured data — send a PDF, image, scanned page, or DOCX file and get back clean, typed JSON. Capture renders any URL or HTML string into a pixel-perfect image or PDF. Both share one API key and one pool of credits.
What you can do
- Extract any document — invoices, receipts, bank statements, tax forms, IDs, and more, using built-in document types or your own custom types.
- Auto-detect the document type when you don't know it in advance.
- Bring your own schema — define the exact fields you want and dOCR returns exactly those keys.
- Capture screenshots — render a URL or raw HTML to PNG, JPEG, WebP, or PDF, with device presets, full-page capture, caching, and embeddable signed URLs.
- Automate end to end — receive signed webhooks when extractions or renders complete.
How it works
Extract — upload a document to POST /api/v1/extract. dOCR parses digital
PDFs and DOCX as text and uses a vision model for images and scanned pages, then
returns structured JSON with a confidence score.
Capture — send a url or html to POST /api/v1/screenshots. dOCR renders
the page in a headless browser and returns a hosted image or PDF.
Supported files & limits
| Extraction formats | .pdf, .jpg, .jpeg, .png, .bmp, .webp, .docx |
| Max size | 10 MB per document |
| Max pages | 15 pages per document |
| Screenshot formats | png, jpeg, webp, pdf |