PDF to markdown, page by page.

Send a PDF or image, pick a frontier model, and get markdown back for each page. You pay the model's token price plus a small premium.

Get an API key Read the docs

curl https://www.opendocrouter.ai/v1/parse \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "google/gemini-3.8-flash-low",
    "document": { "url": "https://arxiv.org/pdf/1706.03762" },
    "pages": "1-3"
  }'

Parallel page parsing

Each page is its own model call. Pages run in parallel as provider capacity becomes available, with a result and status for each page.

An API shaped for parsing

File in, markdown per page out, with a status and a charge for every page. Retry just the pages that failed.

Versioned parsing recipes

Prompts and settings from ParseBench, versioned per model. No parsing prompts of your own to write or maintain.

Prepaid, per token

Top up from $10. Each request is charged for the tokens its pages used. Failed pages are free.

See models and prices