Documents and images to markdown
OCR
Extract text, tables and layout from PDFs and images as clean markdown, one entry per page, with the /v1/ocr endpoint.
POST /v1/ocr
# Invoice 2041| Item | Qty | Total || ------- | --- | ----- || Hosting | 1 | 240 || Support | 2 | 180 |
01
Markdown out
Headings and tables preserved.
02
Per page
One clean entry per page.
03
PDFs & images
By URL or inline data.
04
Dedicated endpoint
Built for OCR models.
Built for
RAG ingestion
Index scanned documents
Invoices
Tables to structured data
Archives
Make old scans searchable
Read a PDF
terminal
1curl https://api.llmgateway.io/v1/ocr \2 -H "Authorization: Bearer $LLM_GATEWAY_API_KEY" \3 -H "Content-Type: application/json" \4 -d '{5 "model": "mistral-ocr-latest",6 "document": {7 "type": "document_url",8 "document_url": "https://arxiv.org/pdf/2201.04234"9 }10 }'