pardesline — Claude Code MCP server: what it does, how to install
pardesline — an MCP server for Claude Code, Codex, and Cursor that connects AI agents to 3D vision and medical imaging datasets.
Apache-2.0 ★ 10 updated 2026-11-16
claude mcp add --env PARDESLINE_CACHE=<PARDESLINE_CACHE> --env HF_TOKEN=<HF_TOKEN> --transport stdio pardesline -- uvx pardeslineWhat pardesline does
The server gives assistants programmatic access to 43 public datasets and 73 algorithms for point clouds, meshes, and CT or MRI volumes. It normalises file formats, spatial frames, and metadata so models can inspect scans and hand results over to Open3D, PyVista, or SimpleITK. Python 3.13 setups require Python 3.12 or the main branch until Open3D compatibility settles.
Environment variables PARDESLINE_CACHE and HF_TOKEN are passed to configure caching and Hugging Face access via uvx. The full Python package installs with pip install "pardesline[all]".
What it helps with
- load a 3D or medical dataset in a script
- perform rigid registration between CT and MRI volumes
- convert data structures between Open3D and PyVista
- run point cloud algorithms on real scans
Example prompts for the agent
- Load the
rire_braindataset and align the MR scan to the CT - Find a registration pipeline for head scans and run it
- Fetch a public 3D dataset through pardesline
How to install pardesline
Platform CLI
claude mcp add --env PARDESLINE_CACHE=<PARDESLINE_CACHE> --env HF_TOKEN=<HF_TOKEN> --transport stdio pardesline -- uvx pardeslineReplace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from pardesline <command>
Platform CLI
codex mcp add pardesline --env PARDESLINE_CACHE=<PARDESLINE_CACHE> --env HF_TOKEN=<HF_TOKEN> -- uvx pardeslineReplace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from pardesline <command>
Config file
{
"mcpServers": {
"pardesline": {
"command": "uvx",
"args": [
"pardesline"
],
"env": {
"PARDESLINE_CACHE": "<PARDESLINE_CACHE>",
"HF_TOKEN": "<HF_TOKEN>"
}
}
}
}Replace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from pardesline <command>
Config file
{
"servers": {
"pardesline": {
"type": "stdio",
"command": "uvx",
"args": [
"pardesline"
],
"env": {
"PARDESLINE_CACHE": "<PARDESLINE_CACHE>",
"HF_TOKEN": "<HF_TOKEN>"
}
}
}
}Replace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from pardesline <command>
Platform CLI
gemini mcp add -e PARDESLINE_CACHE=<PARDESLINE_CACHE> -e HF_TOKEN=<HF_TOKEN> pardesline uvx pardeslineReplace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from pardesline <command>
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 pardesline
How to install pardesline in Claude Code?
Run claude mcp add pardesline -e HF_TOKEN=<HF_TOKEN> -- uvx pardesline in your terminal.
What does pardesline do?
It bridges AI agents with a library of 3D vision and medical imaging datasets and processing algorithms.
What is needed to run pardesline?
You need uvx installed along with the PARDESLINE_CACHE and HF_TOKEN environment variables.