julia-mcp

MCP server that gives AI assistants access to efficient Julia code execution. Avoids Julia's startup and compilation costs by keeping sessions alive across calls, and persists state (variables, functions, loaded packages) between them — so each iteration is fast.

• Sessions start on demand, persist state between calls, and recover from crashes — no manual management

• Each project directory gets its own isolated Julia process

• Pure stdio transport — no open ports or sockets

Tools

• julia_eval(code, env_path?, timeout?) — execute Julia code in a persistent session. env_path sets the Julia project directory (omit for a temporary session). timeout defaults to 60s and is auto-disabled for Pkg operations.

• julia_restart(env_path?) — restart a session, clearing all state. If env_path is omitted, restarts the temporary session.

• julia_list_sessions — list active sessions and their status

Related MCP server: HOPX MCP Server

Requirements

• uv (you might already have it installed)

• Julia – any version, julia binary must be in PATH

  • Recommended packages – used automatically if available in the global environment:

  • Revise.jl - to pick code changes up without restarting

  • TestEnv.jl — to properly activate test environment when env_path points to /test/

The server itself is written in Python since the Python MCP protocol implementation is very mature.

Usage

First, clone the repository:

cd /any_directory
git clone https://github.com/aplavin/julia-mcp.git

Then register the server with your client of choice (see below).

That's it! Your AI assistant can now execute Julia code more efficiently, saving of TTFX.

Claude Code

User-wide (recommended — makes Julia available in all projects):

claude mcp add --scope user julia -- uv run --directory /any_directory/julia-mcp python server.py

Project-scoped (only available in the current project):

claude mcp add --scope project julia -- uv run --directory /any_directory/julia-mcp python server.py

Append Julia flags after server.py to override the defaults (--startup-file=no --threads=auto):

claude mcp add --scope user julia -- uv run --directory /any_directory/julia-mcp python server.py --threads=1 --startup-file=yes

Claude Desktop

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "julia": {
      "command": "uv",
      "args": ["run", "--directory", "/any_directory/julia-mcp", "python", "server.py"]
    }
  }
}

Append Julia flags after server.py to override the defaults (--startup-file=no --threads=auto):

{
  "mcpServers": {
    "julia": {
      "command": "uv",
      "args": ["run", "--directory", "/any_directory/julia-mcp", "python", "server.py", "--threads=1", "--startup-file=yes"]
    }
  }
}

Codex CLI

User-wide — makes Julia available in all projects:

codex mcp add julia -- uv run --directory /any_directory/julia-mcp server.py

Append Julia flags after server.py to override the defaults (--startup-file=no --threads=auto):

codex mcp add julia -- uv run --directory /any_directory/julia-mcp server.py --threads=1 --startup-file=yes

VS Code Copilot

Add to .vscode/mcp.json:

{
  "servers": {
    "julia": {
      "command": "uv",
      "args": ["run", "--directory", "/path/to/julia-mcp", "python", "server.py"]
    }
  }
}

Append Julia flags after server.py to override the defaults (--startup-file=no --threads=auto):

{
  "servers": {
    "julia": {
      "command": "uv",
      "args": ["run", "--directory", "/path/to/julia-mcp", "python", "server.py", "--threads=1", "--startup-file=yes"]
    }
  }
}

GitHub Copilot CLI

Edit $HOME/.copilot/mcp-config.json, and enter

{
  "mcpServers": {
    "julia": {
      "type": "stdio",
      "command": "uv",
      "args": ["run", "--directory", "/path/to/julia-mcp", "python", "-u", "server.py"]
    }
  }
}

Beware that on Windows, \ must be escaped (so write as C:\\my_folder\\...)

GitHub Copilot Cloud Agent

To enable the MCP for a single repo, go to Settings, then scroll down the left panel until you get to Copilot, open that dropdown and select Cloud agent. Then scroll down to the section Model Context Protocol (MCP) and add the following

{
  "mcpServers": {
    "julia": {
      "type": "local",
      "command": "uvx",
      "args": [
        "--from",
        "git+https://github.com/aplavin/julia-mcp",
        "julia-mcp"
      ],
      "tools": ["*"]
    }
  }
}

Details

• Each unique env_path gets its own isolated Julia session. Omitting env_path uses a temporary session that is cleaned up on MCP shutdown.

• If env_path ends in /test/, the parent directory is used as the project and TestEnv is activated automatically. For this to work, TestEnv must be installed in the base environment.

• Julia is launched with --threads=auto and --startup-file=no by default. Pass custom Julia CLI flags after server.py to override these defaults entirely.

Alternatives

Other projects that give AI agents access to Julia:

• MCPRepl.jl and REPLicant.jl require you to manually start and manage Julia sessions. julia-mcp handles this automatically.

• DaemonConductor.jl (linux only) runs Julia scripts, but calls are independent and don't share variables. julia-mcp retains state between calls.