NAPIER

     ●     ●    
       ╳       
     ●     ●    

 N A P I E R
 
 power to people

Overview

NAPIER is a local LLM agent with MCP (Model Context Protocol) capabilities. It allows you to:

1. Chat with local LLM models through Ollama
2. Act as an MCP host that can connect to various MCP-compatible tools
3. Configure and manage these tools through a simple CLI interface

Features

• Local LLM Access: Uses Ollama to run language models locally on your machine
• MCP Host: Acts as a host for the Model Context Protocol
• Tool Integration: Connect to any MCP-compatible tools (like WhatsApp, file systems, web browsers, etc.)
• Easy Configuration: Simple CLI interface to manage your tools and preferences
• API Server: Exposes an API for other applications to interact with NAPIER

Installation

Prerequisites

• Python 3.8+
• Ollama (will be installed automatically if not present)

Setting Up

1. Clone the repository:

   git clone https://github.com/yourusername/napier.git
   cd napier
   

2. Install dependencies:

   pip install -r requirements.txt
   

3. Run NAPIER:

   python napier_cli.py
   

Usage

Chat with Local LLM

NAPIER automatically manages Ollama for you. Once started, you can chat with your local LLM models through a simple interface.

Managing MCP Tools

You can add, remove, and configure MCP tools through the interactive menu:

1. Add Tool: Configure a new MCP-compatible tool
2. Remove Tool: Remove an existing tool
3. Start/Restart Tools: Ensure all configured tools are running

Example: Adding WhatsApp MCP

Tool ID: whatsapp-mcp
Tool Name: WhatsApp MCP
Tool URL: http://localhost:3000
Start Command: npm run start
Command Directory: ./mcp-servers/whatsapp-mcp/
Capabilities: messaging,contact_management
Installation Command: npm install

Configuration

NAPIER stores its configuration in config/napier_config.json. This file contains:

• Basic NAPIER settings
• Default LLM model
• MCP host configuration
• Configured tools

Model Context Protocol (MCP)

The Model Context Protocol is a specification that standardizes how AI assistants interact with local tools and applications. NAPIER implements the MCP Host API, allowing it to:

1. Discover MCP-compatible tools
2. Execute actions on these tools
3. Receive data from tools
4. Manage permissions and security

Learn more about MCP at modelcontextprotocol.io

Developing MCP Tools

You can create your own MCP-compatible tools that work with NAPIER. An MCP tool should:

1. Expose a REST API (typically on localhost)
2. Implement the required MCP endpoints:
   • /status - Returns the status of the tool
   • /capabilities - Lists the capabilities of the tool
   • /actions/{action} - Endpoints for each action the tool can perform

API Reference

NAPIER exposes an API server at http://0.0.0.0:8000 with the following endpoints:

• GET /tools - List all configured tools
• GET /tools/{tool_id} - Get information about a specific tool
• POST /tools/{tool_id}/start - Start a specific tool
• POST /chat - Send a chat message to the local LLM

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License - see the LICENSE file for details.