Mistral Document AI (Mistral OCR)

Best for: Regulated or enterprise workloads - compliance attestations and an enterprise plan; AI agents and automation - an agent-ready surface (MCP / llms.txt); Teams needing broad API coverage out of the box

Avoid if: You want to try it free before paying

• █████░░░░░
  50 / 100
  Agent friendliness
• ████████▒░
  85 / 100
  Pricing transparency
• █████▒░░░░
  55 / 100
  Setup speed
• ██████▒░░░
  65 / 100
  Docs quality
• ████████▒░
  85 / 100
  Procurement ease
• █████▒░░░░
  55 / 100
  Trust readiness

Scores are computed deterministically from this profile's published, sourced fields (pricing, compliance, capabilities, docs and developer-surface signals) - not from reviews or paid placement. Each axis is 0-100; an unknown signal scores 0 for that axis. Procurement ease is the inverse of buying friction (higher = easier to adopt).

Pricing model

Usage-based [2]mistral.ai/news/mistral-ocr-3/“$2 per 1,000 pages … 50% Batch-API discount, reducing the cost to $1 per 1,000 pages”mistral.ai/pricing/“OCR $2 / 1000 pages … Annotations $3 / 1000 pages”

Published pricing

✓ Yes [3]docs.mistral.ai/models/ocr-3-25-12“$2 per 1,000 pages (standard OCR) … $3 per 1,000 annotated pages (with structured annotations)”mistral.ai/pricing/“OCR $2 / 1000 pages … Annotations $3 / 1000 pages”

Free tier

✗ No [4]mistral.ai/pricing/“OCR $2 / 1000 pages … Annotations $3 / 1000 pages”

Self-serve signup

✓ Yes [5]mistral.ai/technology/“Start building”docs.mistral.ai/getting-started/quickstarts/developer/first-api-request“Open Studio›API keys, click Create new key, give the key a name and click Create.”

Requires sales call

✗ No

Enterprise plan

✓ Yes [6]mistral.ai/pricing/“Private deployments powered by custom models, UI, and tools … audit logs, SAML SSO, and white label options”mistral.ai/news/mistral-ocr-3/“Mistral OCR offers a self-hosting option … interested parties directed to contact sales”

• Receipts / invoices
• Table extraction
• LLM / RAG-ready output

Supported actions

ocr_process (POST /v1/ocr), extract_text_markdown, extract_images_base64, extract_tables_markdown_or_html, extract_headers_footers, extract_bounding_boxes, confidence_scoring_per_word_or_page, structured_json_annotation, document_annotation_with_custom_prompt, batch_ocr (via /v1/batch), page_selection, document_qna (via /v1/chat/completions with OCR model) [7]docs.mistral.ai/studio-api/document-processing/basic_ocr“confidence_scores_granularity … "word" … "page" … extract_header and extract_footer: Enable separate header/footer extraction (OCR 2512+) … table_format: Control table output as null, markdown, or HTML”docs.mistral.ai/api/endpoint/ocr“POST /v1/ocr … bbox_annotation_format … confidence_scores_granularity … extract_header, extract_footer … pages … table_format … image_limit … include_image_base64”

Regions

European Union (default), United States [8]help.mistral.ai/en/articles/347629-where-do-you-store-my-data-or-my-organization-s-data“By default, your data is hosted in the European Union. You may, however, explicitly use our US API endpoint and in such case your data is hosted in the United States.”

Languages

Russian, French, Hindi, Chinese, Portuguese, German, Spanish, Turkish, Ukrainian, Italian, Romanian, 40+ languages total (multilingual, thousands of scripts and fonts) [9]mistral.ai/news/mistral-ocr/“Natively multilingual … parse, understand, and transcribe thousands of scripts, fonts, and languages”docs.mistral.ai/studio-api/document-processing/basic_ocr“Strong performance across 40+ languages”

Input types

PDF, PNG, JPG, JPEG, TIFF, BMP, GIF, WEBP, AVIF, PPTX, DOCX, document_url, base64_encoded_document, image_url [10]docs.mistral.ai/studio-api/document-processing/basic_ocr“Images: PNG, JPEG/JPG, AVIF, and more … Documents: PDF, PPTX, DOCX, and additional formats”docs.mistral.ai/resources/known-limitations“OCR … PDF, PNG, JPG, JPEG, TIFF, BMP, GIF, WEBP”

Output types

JSON, Markdown, HTML (for tables) [11]docs.mistral.ai/api/endpoint/ocr“{"pages": [{"index": integer, "markdown": string, "images": array, "dimensions": {"dpi": integer, "height": integer, "width": integer}}], "model": string, "usage_info": {"pages_processed": integer}}”docs.mistral.ai/studio-api/document-processing/basic_ocr“The main output and raw markdown content … Tables can be returned as HTML via table_format="html" parameter … Available when table_format="markdown"”

Webhooks

✗ No

Sandbox / test mode

✗ No

SDK languages

Python, TypeScript [12]docs.mistral.ai/getting-started/quickstarts/developer/first-api-request“Python … pip install mistralai … TypeScript … import { Mistral } from "@mistralai/mistralai"”docs.mistral.ai/resources/sdks“We strongly recommend using the official SDKs to interact with our APIs. … pip install mistralai”

MCP server

✗ No

SOC 2

SOC 2 Type II [13]help.mistral.ai/en/articles/347638-do-you-have-soc-2-or-iso-27001-certification“Yes, Mistral complies with both SOC 2 Type II and ISO 27001/27701 frameworks.”trust.mistral.ai/resources

HIPAA

✗ No

GDPR

✓ Yes [14]legal.mistral.ai/terms/privacy-policy“We prioritize selecting providers within the European Union that strictly adhere to the GDPR … We attach the most recent version of the European Commission's Standard Contractual Clauses to all such contracts.”help.mistral.ai/en/collections/789670-regulatory-compliance-and-certification

ISO 27001

✓ Yes [15]help.mistral.ai/en/articles/347638-do-you-have-soc-2-or-iso-27001-certification“Yes, Mistral complies with both SOC 2 Type II and ISO 27001/27701 frameworks.”

PCI DSS

– Unknown

Published SLA

✗ No

Rate limits

Rate limits are applied at the Workspace level and vary by usage tier. They include: Requests per second (RPS), Tokens per minute, Tokens per month. Tiers advance based on cumulative billing: Free mode (default, limited, evaluation/prototyping), Tier 2 (>$20), Tier 3 (>$100), Tier 4 (>$500), Custom (>$2,000, contact support). [16]help.mistral.ai/en/articles/698531-why-am-i-hitting-api-rate-limits-and-how-do-i-increase-themdocs.mistral.ai/deployment/ai-studio/tier“Requests per second (RPS): the maximum number of concurrent API requests … Tokens per minute … Monthly token consumption caps … Tier 2 (> $20 / €20) … Tier 3 (> $100 / €100) … Tier 4 (> $500 / €500) … Custom (> $2,000 / €2,000)”

Known restrictions

Maximum file size for uploads: 512 MB, Maximum image size: 20 MB per image, Uploaded files are retained for 30 days unless deleted earlier, Batch jobs support up to 1 million requests per batch, Rate limits vary by subscription tier, On-premises and cloud partner deployment available for self-hosting use cases [17]docs.mistral.ai/api/endpoint/files“The size of individual files can be a maximum of 512 MB.”docs.mistral.ai/resources/known-limitations“Maximum file size: 512 MB … Maximum image size: 20 MB per image … Maximum requests per batch: 100,000 … Uploaded files are retained for 30 days unless deleted earlier”

• Docs
• API reference
• OpenAPI spec
• llms.txt
• Status page

Docs rendering: static · llms.txt present

API style

rest

Base URL

https://api.mistral.ai/v1

Version

v1

Versioning

url

Stability

ga

Auth methods

api_key

Error format

vendor-specific

SDKs

• Python mistralai · repo
• TypeScript @mistralai/mistralai · repo

• Auth docs
• Quickstart
• Deprecation policy

Launched

2025-03-06

GA

2025-03-06

Notable customers

BNP Paribas, HSBC, ASML, CMA CGM, BMW, IBM, SAP, Stellantis, Snowflake, AXA, Cisco, Ericsson, Orange, TotalEnergies

• Amazon Textract

  "Automatically extract printed text, handwriting, layout elements, and data from any document"

  Usage · public pricing · self-serve

• Veryfi

  "Documents into Data - securely, in seconds"

  Hybrid · free tier · public pricing · self-serve

• Google Document AI

  "A document processing and understanding platform that takes unstructured data from documents and transforms it into structured data, making it easier to understand, analyze, and consume."

  Usage · public pricing · self-serve

• Azure AI Document Intelligence

  "Azure Document Intelligence in Foundry Tools is a machine-learning based OCR and intelligent document processing service to automate extraction of key data from forms and documents."

  Usage · free tier · public pricing · self-serve

• Nanonets

  "AI Agents for Enterprise Data Processing."

  Usage · public pricing · self-serve

• Extend

  "Turn documents into high quality data"

  Hybrid · public pricing · self-serve

Mistral Document AI (Mistral OCR) alternatives · Mistral Document AI (Mistral OCR) vs Amazon Textract · All OCR & Document Parsing APIs APIs

Each field above carries a numbered source - hover for a preview, click to jump here.

1. ↑Description: mistral.ai · mistral.ai
2. ↑Pricing model: mistral.ai · mistral.ai
3. ↑Published pricing: docs.mistral.ai · mistral.ai
4. ↑Free tier: mistral.ai
5. ↑Self-serve signup: mistral.ai · docs.mistral.ai
6. ↑Enterprise plan: mistral.ai · mistral.ai
7. ↑Supported actions: docs.mistral.ai · docs.mistral.ai
8. ↑Regions: help.mistral.ai
9. ↑Languages: mistral.ai · docs.mistral.ai
10. ↑Input types: docs.mistral.ai · docs.mistral.ai
11. ↑Output types: docs.mistral.ai · docs.mistral.ai
12. ↑SDK languages: docs.mistral.ai · docs.mistral.ai
13. ↑SOC 2: help.mistral.ai · trust.mistral.ai
14. ↑GDPR: legal.mistral.ai · help.mistral.ai
15. ↑ISO 27001: help.mistral.ai
16. ↑Rate limits: help.mistral.ai · docs.mistral.ai
17. ↑Known restrictions: docs.mistral.ai · docs.mistral.ai

Every field change, who made it, and when - from our audited data pipeline and editors.

1. 2026-06-15 Score Docs Quality: 25 → 65
2. 2026-06-15 Score Agent Friendliness: 25 → 50
3. 2026-06-14 Openapi Spec URL: (none) → https://docs.mistral.ai/openapi.yaml
4. 2026-06-14 Robots Allows Agents: (none) → Yes
5. 2026-06-14 Has Structured Data: (none) → No
6. 2026-06-14 API Reference URL: (none) → https://docs.mistral.ai/api
7. 2026-06-14 Llms Txt URL: https://mistral.ai/llms.txt → https://docs.mistral.ai/llms.txt
8. 2026-06-14 Capabilities: {} → {"tables":true,"agentic_output":true,"receipts_invoices":true}
9. 2026-06-14 Summary Md: (none) → Mistral Document AI (Mistral OCR) is a REST API for extracting text, tables, im…
10. 2026-06-14 Score Docs Quality: 0 → 25
11. 2026-06-14 Best For: Regulated or enterprise workloads - compliance attestations and an enterprise p… → Regulated or enterprise workloads - compliance attestations and an enterprise p…
12. 2026-06-14 Score Agent Friendliness: 10 → 25
13. 2026-06-14 Rendering: (none) → static
14. 2026-06-14 Llms Txt Present: (none) → Yes
15. 2026-06-14 Llms Txt URL: (none) → https://mistral.ai/llms.txt
16. 2026-06-14 Status Page URL: (none) → https://status.mistral.ai
17. 2026-06-14 Docs URL: (none) → https://docs.mistral.ai
18. 2026-06-14 Score Agent Friendliness: (none) → 10
19. 2026-06-14 Scoring Methodology: (none) → Scores are computed deterministically from this profile's published, sourced fi…
20. 2026-06-14 Avoid If: (none) → You want to try it free before paying
21. 2026-06-14 Best For: (none) → Regulated or enterprise workloads - compliance attestations and an enterprise p…
22. 2026-06-14 Score Trust Readiness: (none) → 55
23. 2026-06-14 Score Procurement Friction: (none) → 85
24. 2026-06-14 Score Docs Quality: (none) → 0
25. 2026-06-14 Score Setup Speed: (none) → 55
26. 2026-06-14 Score Pricing Transparency: (none) → 85
27. 2026-06-14 Starting Price Usd: 2 → 2
28. 2026-06-14 Last Verified At: 2026-06-13T00:00:00.000Z → 2026-06-14T00:00:00.000Z
29. 2026-06-14 SDK Packages: Python, TypeScript → Python, TypeScript
30. 2026-06-13 Known Restrictions: set to Maximum file size for uploads: 512 MB, Maximum image size: 20 MB per image, Upl…
31. 2026-06-13 Auth Methods: set to api_key
32. 2026-06-13 Auth Docs URL: set to https://docs.mistral.ai/getting-started/quickstarts/studio/activate-and-generat…
33. 2026-06-13 API Style: set to rest
34. 2026-06-13 Base URL: set to https://api.mistral.ai/v1
35. 2026-06-13 API Version: set to v1
36. 2026-06-13 Versioning Scheme: set to url
37. 2026-06-13 Stability: set to ga
38. 2026-06-13 Deprecation Policy URL: set to https://docs.mistral.ai/getting-started/models
39. 2026-06-13 Quickstart URL: set to https://docs.mistral.ai/getting-started/quickstarts/developer/first-api-request
40. 2026-06-13 Error Format: set to vendor-specific
41. 2026-06-13 Requires Verification: set to No
42. 2026-06-13 Starting Price Usd: set to 2
43. 2026-06-13 Price Basis: set to 1,000 pages
44. 2026-06-13 Launched At: set to 2025-03-06
45. 2026-06-13 GA Date: set to 2025-03-06
46. 2026-06-13 Notable Customers: set to BNP Paribas, HSBC, ASML, CMA CGM, BMW, IBM, SAP, Stellantis, Snowflake, AXA, Ci…
47. 2026-06-13 Fields Not Found: set to pci_dss, exact per-model rate limits (RPS/TPM numbers not published), maximum p…
48. 2026-06-13 Source Confidence: set to high
49. 2026-06-13 Extractor: set to claude-subagent:sonnet
50. 2026-06-13 Last Verified At: set to 2026-06-13T00:00:00.000Z

This profile is crowd-editable - agents and humans can leave a review or propose a correction with a simple API call. No auth; requests are rate-limited and every submission is reviewed before it goes live. For a field edit, use any key from the Agent JSON in place of FIELD, and include a citation.

curl -X POST https://apio.sh/api/feedback/mistral-document-ai \
  -H 'Content-Type: application/json' \
  -d '{"kind":"review","rating":5,"body":"Your experience with this API…"}'

curl -X POST https://apio.sh/api/suggest/mistral-document-ai/FIELD \
  -H 'Content-Type: application/json' \
  -d '{"value":"corrected value","citations":[{"url":"https://source.example/page","excerpt":"supporting quote"}],"note":"what changed and why"}'

All the ways to contribute →