Spiral gives your agent every tool with one account. Browse the tools →

MistralDataAvailable nowReleased 2025-12-18

Mistral OCR 3

mistral-ocr-3

Extract text and tables from documents with Mistral OCR 3 through one Spiral API key. Extracts text, tables and images from PDFs and images into markdown with Mistral OCR 3, with optional structured fields. It runs in the background and returns structured JSON. Pay as you go, with no subscription.

Pricing

$0.0026 per page ($0.0039 with document or bounding-box annotations).

Limits

Runs
In the background
Returns
json
Longest queue wait
1 hour
Run deadline
30 minutes
Longest run, end to end
225 minutes
Largest inline input
1 MB
Largest input file
2 GB
Largest output
5 GB

Input

FieldKindDefaultDescription
document requiredJSON—

Document to run OCR on

pagesJSON—

Specific pages to process. Accepts a list of integers or a string of comma-separated numbers and ranges (e.g. '0,1,2' or '0-5' or '0,2-4'). Page numbers start from 0.

image_limitnumber (whole)—

Max images to extract

table_formatone of
markdown, html
—
extract_footertrue or falsefalse

Extract the page footer into the response's footer field and remove it from the markdown content

extract_headertrue or falsefalse

Extract the page header into the response's header field and remove it from the markdown content

image_min_sizenumber (whole)—

Minimum height and width of image to extract

include_blockstrue or falsetrue

Return paragraph-level bounding boxes for all content blocks in the response

include_image_base64true or false—

Include image URLs in response

bbox_annotation_formatJSON—

Structured output class for extracting useful information from each extracted bounding box / image from document. Only json_schema is valid for this field

document_annotation_formatJSON—

Structured output class for extracting useful information from the entire document. Only json_schema is valid for this field

document_annotation_prompttext—

Optional prompt to guide the model in extracting structured output from the entire document. A document_annotation_format must be provided.

confidence_scores_granularityone of
word, page, block
—

Granularity for confidence scores: 'page' (aggregate only), 'word' (per-word scores), or 'block' (per-block scores). Defaults to None (no confidence scores) to keep response payload small.

Same request in code

curl https://api.spiral.ai/v1/runs \
  -H "Authorization: Bearer $SPIRAL_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{
    "provider": "mistral",
    "model": "mistral-ocr-3",
    "input": {
      "document": {
        "document_url": "<document_url>"
      }
    }
  }'

Frequently asked questions

How much does Mistral OCR 3 cost?

$0.0026 per page ($0.0039 with document or bounding-box annotations). You pay as you go from one Spiral balance, with no subscription, and a failed run is not charged.

How do I use Mistral OCR 3 through an API?

Send `POST /v1/runs` with `"provider": "mistral"`, `"model": "mistral-ocr-3"` and the input fields on this page, using one Spiral API key. It runs in the background: poll the run, or name a webhook to be told when it finishes.

Do I need an account with Mistral to use Mistral OCR 3?

No. Spiral runs it on its own Mistral account and charges your Spiral balance.

Can my AI agent use Mistral OCR 3?

Yes. Agents find and run it through Spiral's MCP server, CLI or agent skill, with the same API key.