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

1 Skipped Deployment

Project	Deployment	Preview	Comments	Updated (UTC)
docs	Skipped			Dec 12, 2025 2:02am

Greptile Overview

Greptile Summary

Fixed Mistral OCR parser to correctly handle wrapped API responses when processing PDFs for knowledge base uploads. The API route wraps Mistral's response in {success: true, output: mistralData}, but the parser was treating this as if the pages array was at the root level, causing PDF parsing to fail.

Key Changes:

• Added conditional logic to unwrap ocrResult.output when it exists and pages is not at root level
• Replaced all references to ocrResult with mistralData to use the unwrapped data consistently
• Removed unnecessary comments to improve code readability

The fix ensures PDF parsing works correctly for knowledge base document processing.

Confidence Score: 5/5

• This PR is safe to merge with minimal risk
• The fix correctly addresses a specific bug where the API response wrapping was not being handled. The logic is sound: it checks if ocrResult.output exists and if pages is not at root level, then unwraps to ocrResult.output. All subsequent references consistently use mistralData instead of ocrResult. The change is surgical and doesn't affect any other functionality. No new dependencies, security issues, or breaking changes introduced.
• No files require special attention

Important Files Changed

File Analysis

Sequence Diagram

sequenceDiagram
    participant Client as Tool Client
    participant Parser as mistralParserTool
    participant API as /api/tools/mistral/parse
    participant Mistral as Mistral OCR API
    
    Client->>Parser: Call with filePath & apiKey
    Parser->>API: POST request
    API->>API: Validate auth & file access
    API->>Mistral: POST to /v1/ocr
    Mistral-->>API: Returns {pages: [...], model: "...", ...}
    API-->>Parser: Wraps in {success: true, output: {...}}
    Parser->>Parser: Check if ocrResult.output exists && !ocrResult.pages
    alt Wrapped response (output exists, no pages at root)
        Parser->>Parser: mistralData = ocrResult.output
    else Direct response (pages at root)
        Parser->>Parser: mistralData = ocrResult
    end
    Parser->>Parser: Extract pages from mistralData.pages
    Parser->>Parser: Process markdown content
    Parser-->>Client: Return {success: true, output: {content, metadata}}