Claude Connect

by grapheneaffiliate

340 downloads
Not rated
GitHub

About

An MCP-compliant external connector server enabling Claude and other LLMs to interact with web APIs, filesystem, and custom resources

Details

Author
grapheneaffiliate
Downloads
340
Categories
AI

- Web search integration (Bing required, Google optional)
- File operations within a secure sandbox
- Custom resources and prompt templates
- Multiple transport options (STDIO and HTTP with JWT)
- Extensible, modular architecture
- Cross‑platform (tested on Windows 10, Python 3.10+)

Setting up with Highlight

This MCP is not yet compatible with Highlight’s one-click setup. However, you can still use it with Highlight by following these steps:

  1. Download and install Highlight from highlightai.com/download
  2. Navigate to the plugins tab and select "Add Custom Plugin"
  3. Configure the plugin with the settings below
    Plugin Name Claude Connect
    Command (node, npx, python, etc.)

    Please refer to the README for specific instructions on how to obtain API keys or other required environment variables.

  4. Enable "Start Automatically" if you want the plugin to start when Highlight launches

From the repository

Install Python 3.10+, create and activate a virtual environment, install dependencies (..\venv\Scripts\pip.exe install -r requirements.txt), rename .env.example to .env and add your API keys (Bing Search API required, Google Search API optional). Run the server in STDIO mode (..\venv\Scripts\python.exe main.py) or HTTP mode (..\venv\Scripts\python.exe main.py --http). Test with ..\venv\Scripts\pytest.exe. For Claude Desktop integration, follow the dedicated guide.

Claude Desktop / Cursor

Paste into your MCP client config file to install this server.

{
    "mcpServers": {
        "claude connect": {
            "claude-connect": {
                "command": "docker",
                "args": [
                    "build",
                    "-t",
                    "claude-connect",
                    "."
                ]
            }
        }
    }
}

McpServers

{
    "claude-connect": {
        "command": "docker",
        "args": [
            "build",
            "-t",
            "claude-connect",
            "."
        ]
    }
}

Claude Connect

An MCP-compliant external connector server enabling Claude and other LLMs to interact with web APIs, filesystem, and custom resources. v1.1 (March 2025)

Overview

Claude Connect is a server that implements Anthropic's Model Context Protocol (MCP), allowing Large Language Models like Claude to interact with external tools and resources in a standardized way. This server enables LLMs to access web search results, read and write files in a sandboxed environment, and utilize custom resources.

Key Features

- Web Search Integration: Connect to Bing Search API (required) and Google Search API (optional)
- File Operations: Read, write, list, and delete files within a secure sandbox
- Custom Resources: Define and access custom resource types
- Prompt Templates: Store and retrieve custom prompt templates
- Multiple Transport Options: Primary STDIO transport for Claude Desktop integration, optional HTTP transport with JWT authentication
- Extensible Architecture: Easy to add new capabilities and resources
- Cross-Platform: Tested on Windows 10, but designed to work on any platform supporting Python 3.10+

Quick Start

This MCP server enables Large Language Models like Claude to access external tools and resources. It takes about 15 minutes to set up and requires Python 3.10+ and API keys for search functionality. Follow these steps to get started:

1. Install Python 3.10+, create a virtual environment, and activate it (e.g., using python -m venv venv and ..\venv\Scripts\activate on Windows).
2. Install dependencies: ..\venv\Scripts\pip.exe install -r requirements.txt (using full path helps bypass potential PowerShell execution policy issues).
3. Rename .env.example to .env and add your API keys (Bing Search API required, Google Search API optional).
4. Run server in STDIO mode: ..\venv\Scripts\python.exe main.py or HTTP mode: ..\venv\Scripts\python.exe main.py --http.
5. Test with ..\venv\Scripts\pytest.exe.

Note: Some commands, especially activation or installation, might require running PowerShell/CMD as Administrator on Windows depending on your system configuration.

For detailed setup instructions, see the Setup Guide. For Claude Desktop integration, see the Claude Desktop Integration Guide.

Documentation

- Setup Guide
- Configuration Guide
- Claude Desktop Integration
- API Reference
- Advanced Usage
- Troubleshooting

Advanced Usage

For advanced usage scenarios, including:

- Manual JSON-RPC interaction
- Extending the server with new capabilities
- Using PostgreSQL instead of SQLite
- Enabling Redis caching
- Internationalization
- Performance tuning

See the Advanced Usage Guide.

Code Explanation & Design Rationale

The Claude Connect server is designed with the following principles in mind:

- Modularity: Each capability is implemented as a separate handler function.
- Security: File operations are restricted to a sandbox directory, and path traversal attempts are prevented.
- Extensibility: New capabilities can be added by implementing new handler functions and updating the capabilities list.
- Configurability: Most aspects of the server can be configured through environment variables and configuration files.

The server supports two transport modes:

- STDIO: For integration with Claude Desktop and other MCP clients that support STDIO transport.
- HTTP: For development, testing, and integration with web applications.

For more details on the code structure and design decisions, see the API Reference.

Containerization & Deployment

You can containerize Claude Connect using Docker:

```dockerfile

No reviews yet — be the first

Sign in to leave a review

Use Google, GitHub, or an email account so ratings stay tied to real people.

Email sign in

No reviews posted yet.