mcp · hushvert — Claude Code MCP server: what it does, how to install
Converts office files, PDF documents, and video through the hushvert cloud API in AI coding setups. Useful when your machine lacks LibreOffice or local conversion packages.
claude mcp add --env HUSHVERT_API_KEY=<HUSHVERT_API_KEY> --env HUSHVERT_ALLOWED_DIR=<HUSHVERT_ALLOWED_DIR> --transport stdio hushvert-mcp -- npx -y @hushvert/mcpWhat mcp does
The server connects coding agents to the hushvert hosted API over the Model Context Protocol. Its primary tool, convert_file, sends a local document to the cloud service and saves the output right next to the source file. For long operations like large video transcoding, the tool yields a jobId after wait_seconds, which defaults to 120. The agent checks progress via convert_poll.
It handles conversions outside browser capabilities. Supported pairs include DOCX or PPTX to PDF, PDF to Word, and video containers like MOV or MKV to MP4. The server refuses audio or image conversion, directing agents to @hushvert/engine instead.
With LibreOffice installed, skip this server. Agents on bare machines often fall back to pandoc. Pandoc rebuilds files through LaTeX, discarding heading colors and original fonts. This server avoids installing heavy office suites in CI containers, though it requires HUSHVERT_API_KEY.
What it helps with
- convert DOCX and PPTX files to PDF while preserving layout styling
- turn PDF documents into editable DOCX files
- transcode MKV and MOV video files to MP4
- run office conversions in CI runners without installing LibreOffice
Example prompts for the agent
- Convert report.docx to PDF with hushvert
- Turn slides.pptx into PDF and place the output file next to it
- Transcode archive.mkv to mp4
How to install mcp
Platform CLI
claude mcp add --env HUSHVERT_API_KEY=<HUSHVERT_API_KEY> --env HUSHVERT_ALLOWED_DIR=<HUSHVERT_ALLOWED_DIR> --transport stdio hushvert-mcp -- npx -y @hushvert/mcpReplace <…> with your own values
Platform CLI
codex mcp add hushvert-mcp --env HUSHVERT_API_KEY=<HUSHVERT_API_KEY> --env HUSHVERT_ALLOWED_DIR=<HUSHVERT_ALLOWED_DIR> -- npx -y @hushvert/mcpReplace <…> with your own values
Config file
{
"mcpServers": {
"hushvert-mcp": {
"command": "npx",
"args": [
"-y",
"@hushvert/mcp"
],
"env": {
"HUSHVERT_API_KEY": "<HUSHVERT_API_KEY>",
"HUSHVERT_ALLOWED_DIR": "<HUSHVERT_ALLOWED_DIR>"
}
}
}
}Replace <…> with your own values
Config file
{
"servers": {
"hushvert-mcp": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@hushvert/mcp"
],
"env": {
"HUSHVERT_API_KEY": "<HUSHVERT_API_KEY>",
"HUSHVERT_ALLOWED_DIR": "<HUSHVERT_ALLOWED_DIR>"
}
}
}
}Replace <…> with your own values
Platform CLI
gemini mcp add -e HUSHVERT_API_KEY=<HUSHVERT_API_KEY> -e HUSHVERT_ALLOWED_DIR=<HUSHVERT_ALLOWED_DIR> hushvert-mcp npx -y @hushvert/mcpReplace <…> with your own values
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about mcp
How do I install mcp in Claude Code?
Run claude mcp add mcp -e HUSHVERT_API_KEY=<HUSHVERT_API_KEY> -- npx -y @hushvert/mcp in your terminal. Replace <HUSHVERT_API_KEY> with your active key from hushvert.
What does mcp do?
It gives coding agents the convert_file tool to process office documents, PDFs, and video files through the hosted hushvert API. The converted file is saved next to the input path.
Why use this server instead of pandoc for Word documents?
Pandoc recreates documents from an internal syntax tree using LaTeX, stripping table styles, header colors, and original fonts like Calibri. Hushvert keeps the source layout intact.
Similar MCP servers
All MCP servers for Office documents →mcp by other authors comes first.
