AI Agent Integrations: Client Configurations

1. Claude Desktop

Claude Desktop communicates with MCP servers to extend Claude's knowledge. Add the configuration block below to your Claude Desktop config file:

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "pyproc": {
      "command": "pyproc-mcp",
      "env": {
        "PYPROC_TIMEOUT": "30"
      }
    }
  }
}

2. Cursor

Cursor supports custom local command-based MCP server endpoints.

To register PyProc in Cursor, navigate to Settings → Features → MCP:

1. Click + Add New MCP Server.
2. Set Name to pyproc.
3. Set Type to command.
4. Set Command to: pyproc-mcp.

3. Claude Code

Claude Code resolves local MCP servers using standard process launching.

Installation Command:

claude mcp add pyproc -- pyproc-mcp

Manual Config Scope:

Alternatively, edit your global user config file at ~/.claude.json or project-local config at .claude/mcp.json:

{
  "mcpServers": {
    "pyproc": {
      "command": "pyproc-mcp"
    }
  }
}

4. Google Antigravity

Antigravity registers external MCP tooling dynamically inside its modular sidecar engine.

Manual Configuration:

Add the server configuration payload to your global config file at ~/.gemini/config/mcp_config.json or project-local configuration directory at .agents/mcp_config.json:

{
  "mcpServers": {
    "pyproc": {
      "command": "pyproc-mcp"
    }
  }
}

5. Codex

Codex manages tool execution lists inside TOML configurations.

Installation Command:

codex mcp add pyproc -- pyproc-mcp

TOML Configuration:

Add to your settings at ~/.codex/config.toml or local .codex/config.toml:

[mcp_servers.pyproc]
command = "pyproc-mcp"
args = []

6. OpenCode

OpenCode loads servers using custom configurations referencing local execution blocks.

JSON Configuration:

Add to your settings file at ~/.config/opencode/opencode.json or project root opencode.json:

{
  "mcp": {
    "pyproc": {
      "type": "local",
      "command": ["pyproc-mcp"],
      "enabled": true
    }
  }
}

7. OpenClaw

OpenClaw controls standard local tools processes via commands setup and gateways.

Configuration Commands:

# Register the pyproc server command
openclaw mcp set pyproc '{"command":"pyproc-mcp","args":[]}'

# Restart the local openclaw gateway
openclaw gateway restart

8. Hermes Agent

Hermes Agent is configured via a global YAML configuration file.

Configuration Settings:

Modify your config file at ~/.hermes/config.yaml by appending the command setup block:

mcp_servers:
  pyproc:
    command: "pyproc-mcp"

9. Generic MCP Tutorial

Any standard MCP client that executes external processes via command interfaces can integrate with PyProc. Ensure pyproc-mcp is globally accessible in your system shell path (e.g. within active virtual environments).

For custom configurations, pass environment variables alongside execution parameters:

• Command: pyproc-mcp
• Environment:
  • PYPROC_TIMEOUT: Request timeout in seconds.
  • PYPROC_SSL_VERIFY: Set to 1 or true to enforce SSL checks.