ParaView MCP

Connect ParaView to LLM assistants through the Model Context Protocol.

The Python server is built with FastMCP 3.x. Support for the 2026-07-28 MCP specification is planned once FastMCP 4 reaches a stable release.

paraview-mcp-server has two runtime parts:

• a ParaView plugin (C++/Qt) that exposes a TCP bridge inside the ParaView GUI
• a Python MCP server that connects to the plugin and serves tools to any MCP client

Prerequisites

• ParaView ≥ 5.13. Pre-built plugins are available for the exact versions and platforms listed below; other ParaView releases require a source build against the matching SDK.
• uv

Quick Start

First set up the ParaView plugin. Then add the Python MCP server to Claude Code in one command:

claude mcp add paraview -- uvx paraview-mcp-server

Open Tools > ParaView MCP in ParaView, start the bridge, and connect from Claude Code.

Set Up the ParaView Plugin

Download a pre-built plugin binary from the latest GitHub Release. Releases provide this matrix:

Choose the package that names your exact ParaView version and platform. Download its adjacent .sha256 file, verify the package, then open or extract it and follow the included INSTALL.md. Pull requests also produce corresponding platform binaries as short-lived GitHub Actions artifacts; GitHub Releases are the permanent distribution channel.

macOS release images are Developer ID-signed, notarized by Apple, and include a stapled notarization ticket. Open the .dmg, copy the contained plugin directory to a persistent location, and load ParaViewMCP.so from that copied directory. Pull-request artifacts are unsigned test builds and remain .tar.gz files.

Alternatively, build the plugin from source against a ParaView 5.13 or newer SDK. See CONTRIBUTING.md for full build instructions. Binary compatibility is release-series specific, so use a plugin built for your ParaView major.minor version.

Once installed:

1. Open Tools > Manage Plugins in ParaView.
2. Click Load New... and select ParaViewMCP.so (Linux/macOS) or ParaViewMCP.dll (Windows) from the plugin directory.
3. Enable Auto Load.
4. Open Tools > ParaView MCP.
5. Click Start Server.

The ParaView MCP panel shows the connection status and execution history. Non-loopback binds require an auth token.

Configure Your MCP Client

Claude Code (CLI)

claude mcp add paraview -- uvx paraview-mcp-server

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "paraview": {
      "command": "uvx",
      "args": ["paraview-mcp-server"]
    }
  }
}

Other MCP Clients

Configure a local stdio MCP server with uvx as the command and paraview-mcp-server as its only argument:

{
  "mcpServers": {
    "paraview": {
      "command": "uvx",
      "args": ["paraview-mcp-server"]
    }
  }
}

Consult your client's documentation for the location and exact format of its MCP server configuration.

Configuration

The server connects to the ParaView plugin using these environment variables:

Defaults work for a standard local setup. Override these when connecting to ParaView on a remote machine or non-standard port:

{
  "mcpServers": {
    "paraview": {
      "command": "uvx",
      "args": ["paraview-mcp-server"],
      "env": {
        "PARAVIEW_HOST": "192.168.1.10",
        "PARAVIEW_PORT": "9877",
        "PARAVIEW_AUTH_TOKEN": "your-token"
      }
    }
  }
}

Available Tools

One command runs at a time, up to three wait in FIFO order, and further calls report busy. This prevents concurrent mutations of ParaView's shared state.

Execution results include output, diagnostics, and separate request and Python statuses. See execution and state for cancellation, timeouts, and recovery behavior.

Design and Differences from ParaView_MCP

This project follows the approach of Blender-MCP and Slicer-MCP, both of which give LLMs direct code execution inside their respective application runtimes.

The existing ParaView_MCP implementation¹ takes a different approach, exposing a fixed set of high-level tools without access to the underlying Python runtime, which limits flexibility for custom workflows. The major differences are:

1. We provide an execute_paraview_code tool that runs arbitrary Python inside the ParaView session. The plugin records each execution and, when ParaView can capture a pipeline snapshot, lets the user restore the state from immediately before that execution. This makes generated scripts easier to inspect, reuse, and adapt for tasks such as batch processing.
2. Architecturally, ParaView_MCP's own disclaimer states that it relies on synchronization between pvserver and the ParaView client. That synchronization mechanism is deprecated in recent ParaView versions and can cause incorrect application views and general stability issues. This project instead runs a plugin inside the interactive ParaView process and exposes a TCP bridge, avoiding the pvserver/client synchronization path entirely.

Contributing

See CONTRIBUTING.md for build instructions, development setup, and pull request guidelines.

License

MIT — see THIRD-PARTY-NOTICES.txt for dependency licenses.