Principle:Mistralai Client python OCR Result Processing

Knowledge Sources	Mistral AI OCR Mistral Client Python
Domains	Document_Processing, OCR
Last Updated	2026-02-15 14:00 GMT

Overview

A response extraction pattern that navigates OCR response structures to retrieve per-page text, tables, images, and document annotations.

Description

OCR Result Processing extracts structured content from the OCRResponse object. The response contains a pages list where each OCRPageObject has markdown (extracted text), images (extracted image objects with optional base64 data), tables (structured table data), and dimensions (page width/height). An optional document_annotation field at the top level contains structured extraction results.

Usage

Use this principle after calling client.ocr.process() to extract and process document content. Iterate over response.pages to access per-page results.

Theoretical Basis

OCR response hierarchical structure:

• OCRResponse → top-level container
  • pages: List of per-page results
    • markdown: Full text content as markdown
    • images: Extracted images with IDs and optional base64 data
    • tables: Structured table objects
    • dimensions: Page width and height
  • usage_info: Token consumption
  • document_annotation: Optional structured extraction

Related Pages

Implemented By

• Implementation:Mistralai_Client_python_OCRResponse_Model