[waleedlatif1]

Summary

• Mistral Parser (v3) gets an Extract Images switch: images Mistral OCR finds in the PDF (figures, logos, photos) come back as stored files on a new images (file[]) output, so an agent can pick the ones it wants and pass them on
• Image bytes never travel through block outputs: each page image keeps its bounding box and gets a file_name matching its stored file, and image_base64 is stripped
• Caps: at most 100 images per parse (image_limit is clamped), a 20 MB per-image cap (bigger images are skipped, not failed), and the shared 100 MB buffered-transfer budget across all images
• New Tables option passes Mistral's table_format (markdown / html), so pages[].tables is populated instead of always empty; default stays inline
• Advanced Max Images / Min Image Size fields, validated with the shared parseOptionalNumberInput
• Only mistral_parser_v3 extracts images; v1/v2 never request them, and the knowledge-base OCR path (executeMistralParse) is unchanged

Type of Change

• New feature

Testing

• New lib/internal/mistral/images.test.ts and 4 new cases in operations.test.ts: base64 stripped and files named, 100-image cap, oversized image skipped, image_limit clamp (above cap / unset / zero), v3-only gate. Each guard was reverted and its test went red
• bunx vitest run lib/internal/mistral tools/mistral blocks — 851 passed
• check-block-registry.ts origin/staging, check:canvas-sentences, check:integration-catalog, check:docs; regenerated tool metadata and integration docs
• lint:check + type-check via the pre-push hook; full check:audits and bun run test left to CI

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing (new tests pass the test-audit authoring gate)
• No new warnings introduced
• I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
docs	Ready	Preview	Nov 10, 2026 10:23pm UTC

[greptile-apps]

Confidence Score: 5/5

[Medium impact] The PR appears safe to merge; no actionable issues remain from this review.

Summary

Adds image files and separate-table options to Mistral Parser v3.

• Mistral Parser v3 returns extracted PDF images as files.
• Mistral Parser v3 can return tables separately as Markdown or HTML.

Diagram

%%{init: {'theme': 'neutral'}}%%
flowchart TD
  A[Mistral Parser v3] --> B[Mistral OCR]
  B --> C{Extract Images enabled?}
  C -->|No| D[Return structured pages]
  C -->|Yes| E[Remove base64 and apply image caps]
  E --> F[Store image files]
  F --> G[Return pages and file references]

Loading

_{Reviews (4) · Last reviewed commit: "docs(mistral): describe Extract Images a..." · Reviewed by Greptile}

[cubic-dev-ai]

All reported issues were addressed across 13 files

<sub>Reply with feedback, questions, or to request a fix.

Turn on auto-fix | Re-trigger cubic</sub>

[waleedlatif1]

@greptile

[waleedlatif1]

@cubic-dev-ai review this PR

[cubic-dev-ai]

@cubic-dev-ai review this PR

@waleedlatif1 I have started the AI code review. It will take a few minutes to complete.

[cubic-dev-ai]

All reported issues were addressed across 14 files

<sub>Reply with feedback, questions, or to request a fix.

Turn on auto-fix | Re-trigger cubic</sub>

[waleedlatif1]

@greptile

[waleedlatif1]

@cubic-dev-ai review this PR

[cubic-dev-ai]

@cubic-dev-ai review this PR

@waleedlatif1 I have started the AI code review. It will take a few minutes to complete.

[cubic-dev-ai]

No issues found across 14 files

Confidence score: 5/5

• Automated review surfaced no issues in the provided summaries.
• No files require special attention.

_{Turn on auto-fix | Re-trigger cubic}