Suppr MCP Server

Document translation in suppr-mcp begins with create_translation, which accepts a local file path or a document URL and a target language. The source language can be specified or detected automatically, and PDF translation includes an optional setting for mathematical formula optimization. Supported formats include PDF, DOCX, PPTX, XLSX, HTML, TXT, and EPUB. After submission, get_translation uses the task ID to report status and progress and provide a download URL when the translated file is ready. For reviewing past work, list_translations returns a paginated task history.

Literature discovery uses search_documents for AI-driven semantic searches with natural-language queries, including Chinese queries against PubMed. Callers can limit the result count, choose which document fields to return, and control automatic result selection. Available fields cover titles, abstracts, authors, DOIs, PubMed IDs, publication names, publication years, and links. Search responses also report consumed points. The service uses stdio transport and a Suppr API key, with the README showing how to connect it to Claude Desktop.