OCR Document
mistral.ocr
ActionAI & modelsconnection required
Calls the Mistral AI API with the connection's credentials.
Run Mistral OCR over a PDF or image and return the text of every page as Markdown, keeping the document's layout and tables.
Connection
- Required
- Yes — the node will not save without one
- Auth type
- API key
- Providers
- Mistral AI
- Secret shape
- {"api_key": "…"} — a Mistral API key
- Config it needs
- base_url, default_model
- Resolved by
- config.connection_id, else node.connection_id
Configuration
6 fields, none required. A dotted path is a nested object and name[] an array of objects — the editor renders the same tree.
| Field | Type | Required | Default | Constraints | Example |
|---|---|---|---|---|---|
document_url URL of the PDF to read. Use a Mistral signed file URL for an uploaded file. | string | No | — | url format | — |
image_url URL of an image to read, when the source is a picture rather than a PDF. | string | No | — | url format | — |
document The raw document chunk as JSON, for shapes the two URL fields do not cover. | object | No | — | — | — |
model OCR model; defaults to mistral-ocr-latest when left blank. | string | No | mistral-ocr-latest | — | — |
pages Zero-based page numbers to process, as a JSON array. Blank means all pages. | array | No | — | array of integer | — |
include_image_base64 Return the extracted images inline as base64 alongside the text. | boolean | No | false | — | — |
- At least one of document_url, image_url, document must be set.
Example configuration
Every value below is the contract’s own example or default — nothing invented.
{
"include_image_base64": false,
"model": "mistral-ocr-latest"
}Output
One shape.
result
| Field | Type | Example |
|---|---|---|
pages One entry per page, as Mistral returned it. | array | — |
pages[].index Zero-based page number. | number | — |
pages[].markdown That page's text as Markdown. | string | — |
page_count How many pages came back. | number | — |
markdown Every page's Markdown joined into one document. | string | — |
model The OCR model that ran. | string | — |
pages_processed Pages Mistral billed for this call. | number | — |
{
"markdown": "",
"model": "",
"page_count": 0,
"pages": [
{
"index": 0,
"markdown": ""
}
],
"pages_processed": 0
}See OCR Document run in a workflow.
In a technical demo we add the Mistral AI connection, drop the node on the canvas and let Workflow Health pre-flight it before the first real run.