What is Agent Cli Mcp On Aws Bedrock

Agent-CLI-MCP-On-AWS-Bedrock, also known as Q-2001, is a command-line AI assistant powered by AWS Bedrock. It utilizes advanced large language models to answer questions, provide assistance, and engage in conversations, integrating intelligent tool capabilities through Model Context Protocol (MCP).

Use cases

Use cases include providing instant customer support responses, assisting in coding tasks, engaging users in educational dialogues, and facilitating complex queries through intelligent tool integrations.

How to use

To use Agent-CLI-MCP-On-AWS-Bedrock, ensure Python 3.11+ is installed, install the required dependencies using ‘pip install -r requirements.txt’, and configure your AWS credentials. You can run the assistant by executing ‘python Q-2001.py’ and specify models, AWS profiles, and regions as needed.

Key features

Key features include a simple command-line interface with rich formatting, support for multiple AWS Bedrock models (like Claude, DeepSeek, Nova), real-time streaming responses, conversation history management, tool usage capabilities via MCP integration, and customizable configurations.

Where to use

Agent-CLI-MCP-On-AWS-Bedrock can be used in various fields such as customer support, educational tools, software development assistance, and any other domain requiring intelligent conversational agents.

Agent CLI MCP on AWS Bedrock (AKA Q-2001)

Q-2001 is a command-line AI assistant powered by AWS Bedrock. It uses advanced large language models to answer questions, provide help, and engage in conversations, with intelligent tool capabilities through Model Context Protocol (MCP) integration.

Features

• Simple and easy-to-use command line interface with rich formatting
• Support for multiple AWS Bedrock models (Claude, DeepSeek, Nova, etc.)
• Real-time streaming responses
• Conversation history and context management
• Tool usage capabilities through MCP integration
• Customizable configuration

Installation

1. Ensure Python 3.11+ is installed
2. Install dependencies:

pip install -r requirements.txt

3. Ensure AWS credentials are configured (via AWS CLI or environment variables)

Usage

# Basic usage
python Q-2001.py

# Specify model
python Q-2001.py --model us.anthropic.claude-3-7-sonnet-20250219-v1:0

# Use other models
python Q-2001.py --model us.deepseek.r1-v1:0
python Q-2001.py --model us.amazon.nova-micro-v1:0

# Specify AWS profile
python Q-2001.py --profile myprofile

# Specify AWS region
python Q-2001.py --region us-east-1

# Show version
python Q-2001.py --version

Commands

The following commands are available in Q-2001:

• /quit or /exit: Exit the application
• /help: Display help information
• /clear: Clear conversation history
• /model <model_id>: Change AI model
• /tools: List all available MCP tools

MCP Integration

Q-2001 supports extended functionality through Model Context Protocol (MCP). The MCP client handles the integration with various tool servers.

Configuration

Q-2001 creates a configuration file at ~/.q2001_mcp_config.json which can be manually edited to change default settings.

Example configuration:

{
  "mcpServers": {
    "aws-document": {
      "command": "uvx",
      "args": [
        "awslabs.aws-documentation-mcp-server@latest"
      ]
    },
    "mcp-server-time": {
      "command": "uvx",
      "args": [
        "mcp-server-time",
        "--local-timezone=America/New_York"
      ]
    },
    "filesystem": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-filesystem",
        "/home/ec2-user/"
      ]
    },
    "sequential-thinking": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-sequential-thinking"
      ]
    }
  }
}

Tool Capabilities

When MCP servers are configured and initialized, Q-2001 can:

1. Automatically detect when tools are needed to solve a problem
2. Call appropriate tools with proper parameters
3. Process tool results and incorporate them into responses
4. Handle multiple tool calls in a single conversation

Requirements

• Python 3.11+
• Required Python packages:
  • boto3
  • rich
  • mcp
  • anyio
  • markdown-it-py
  • uvicorn
  • python-dotenv
  • httpx
  • pygments
• AWS account and appropriate permissions to access Bedrock services
• Configured AWS credentials
• Node.js and npm (for npx commands with certain MCP servers)

Notes

• Using this tool will incur AWS Bedrock API call costs
• Ensure your AWS account has Bedrock service enabled and access to required models