mcp-mistral-ocr
MCP Mistral OCR is a server that utilizes Mistral AI's OCR API to provide functionality for processing images and PDF files for text recognition. It supports both local and URL file processing, and offers convenient features such as Docker containerization and multiple file format support.
What file types are supported by MCP Mistral OCR?
It supports images such as JPG, JPEG, PNG, GIF, WebP, and documents like PDF.
What are the limitations of the Mistral API?
The maximum file size is 50MB and the maximum number of document pages is 1000.
How are OCR results saved?
Results are saved in JSON format in the output directory inside OCR_DIR, with filenames including the original filename and a timestamp.
Can MCP Mistral OCR process files from URLs?
Yes, it can process files from URLs with explicit file type specification.
Is Docker required to run MCP Mistral OCR?
While Docker is recommended for containerization, local development can be done using the UV package manager.