MCP -setup

MCP Chat is a Python-based command-line interface that enables interactive chat with AI models through the Anthropic API. It supports document retrieval by referencing documents in queries and integrates extensible MCP-built commands. You interact with the model by typing messages, include document content using an @document_id tag, and execute specialized prompts via commands prefixed with /. The available tools and commands are defined within the MCP setup and can be extended by editing the mcp_server.py and related client logic. To begin, ensure your Anthropic API key is configured in the environment and run the server using Python as described in the installation steps. The interface will auto-complete commands, making it easier to discover and use the built-in MCP features.

Prerequisites:

• Python 3.9+
• Anthropic API Key

Option A: Setup with uv (Recommended)

1. Install uv:

pip install uv

2. Create and activate a virtual environment:

uv venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

3. Install dependencies in editable mode:

uv pip install -e .

4. Run the project:

uv run main.py

Option B: Setup without uv

1. Create and activate a virtual environment:

python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

2. Install dependencies:

pip install anthropic python-dotenv prompt-toolkit "mcp[cli]==1.8.0"

3. Run the project:

python main.py

Environment setup:

• Create or edit .env in the project root and set ANTHROPIC_API_KEY with your Anthropic secret key. Example: ANTHROPIC_API_KEY="sk-..."

Tips and notes:

• Ensure your Anthropic API key is correctly configured as ANTHROPIC_API_KEY in the environment or in .env.
• If using uv, the installation and run steps are slightly different (use uv commands) but the end goal is to run the Python entry point (main.py).
• Commands and document references follow the MCP pattern; use / to invoke commands and @doc_id to pull in document content.
• The docs dictionary and MCP features may require completing TODOs in mcp_server.py and mcp_client.py to enable full MCP capability.
• If you encounter import or dependency issues, verify your virtual environment is activated and dependencies are installed in the active environment.