$1 per 1,000 pages · first 1,000 free

One API for open-weights
document parsers.

We host open-weights parsing models and normalize what they return, so the output shape stays the same when the parser changes.

Create an API keyRead the docs
PaddleOCR-VL available nowBlock citations on every valueSync, async, or batch
1
2
3
b1 title
PAYMENT AUTHORIZATION
b2 field
Account
4021-8837
b3 field
Date
May 2025
b4 table
ItemAmount
Consulting$4,636.40
Adjustment-$318.20
Total$4,318.20
b5 checkbox
PaidPending
b6 signature
readingpaddleocr-vl-1.6
analyzing
0/6 blocks
/ api

Parse and extract run sync, async, or batch.

Endpoints for parsing, schema extraction, job status, document listing, and file retrieval. Every request uses the same API key.

endpointsapi.openparser.dev
  • Full API reference
POST/parserequest
curl https://api.openparser.dev/parse \
  -H "Authorization: Bearer op_live_…" \
  -F "file=@statement.pdf"
response200 · 1.9s
{
  "documentId": "doc_3f9a2c",
  "filename": "statement.pdf",
  "pageCount": 1,
  "markdown": "## Payment Authorization\n| Item | …",
  "output": {
    "chunks": [
      { "blocks": [
        { "id": "b4", "type": "table",
          "content": "| Item | Amount |\n| …",
          "boundingBox": { "left": 14, "top": 53,
                           "right": 146, "bottom": 123 },
          "metadata": {
            "page": { "number": 1, "width": 1240, "height": 1754 },
            "avgOcrConfidence": 0.98 } }
      ] }
    ]
  },
  "elapsedMs": 1904
}
/ console

Test and manage OpenParser in the console.

Sign in to test requests in the playground, create keys per environment, review page usage and invoices, and manage team access. An open-weights model does the parsing. You pick the LLM that does the extraction.

openparser.dev/playgroundsigned in
playgroundno code
statement.pdf1 page · 48 KB
parsePaddleOCR-VL-1.6open
extractClaudeGPT-4oLlamalocalany LLM
Run{ total: "$4,318.20" } b4
Try the playgroundCreate an API key
/ output

Make your documents agent-ready.

Typed blocks carry ids and per-block OCR confidence. Extraction names those ids instead of inventing coordinates, so an agent cites what it read and an invalid id resolves nowhere.

Typed blocks or markdown

Each page returns typed layout blocks, markdown, or both. Block types include headings, tables, figures, and signatures.

b1
0.99
b4
0.98
b6
0.70

OCR confidence per block

Every parsed block carries an OCR confidence score, so you can set a threshold and route low-confidence pages to review.

One format across parsers

One parser calls it a figure, another an image. Block types map to one schema. PaddleOCR-VL, open weights, is the adapter available today.

schema {
total
}
{
total: "$4,318"
}

Choose the extraction LLM separately

The parsing model does not set the extraction LLM. Point a schema at a parse and fill it with Claude, GPT, or a model you host.

b4

Block citations on every value

Each extracted value carries the id of the block it was read from, so you can highlight that region on the source page.

b2b4b6
x, y, w, h

No generated coordinates

The model names block ids, not geometry. An id that is not in the parse fails to resolve instead of pointing at a region that was never there.

/ pricing

Pay for the pages you parse.

The first 1,000 pages are free. After that, parsing is $1 per 1,000 pages. Extraction adds the token cost of whichever model you pick.

start free1,000pages free
pay per page
$1/ 1,000 pages
async or batch
extractionPer tokenbilled based on model
  • No minimums
  • No per-seat fees
  • Sync, async, and batch
  • Block citations
  • Console and API keys
Create an API key
/ 1,000 pages free

Try it on your
own documents.

Create a key, POST a document, and get back typed blocks, markdown, and JSON with a block citation on every value. The first 1,000 pages are free.

Create an API keyRead the docs