Skip to content

README.md

Latest commit

 

History

History
82 lines (56 loc) · 1.87 KB

README.md

File metadata and controls

82 lines (56 loc) · 1.87 KB
Raw

IBM i MCP Server - Python Client Examples

This directory contains example AI agents that interact with the IBM i MCP Server.

Quick Setup

1. Install uv (recommeded)

curl -LsSf https://astral.sh/uv/install.sh | sh

1. Create Virtual Environment and Install Dependencies

cd client
uv sync

2. Start the MCP Server

Ensure the IBM i MCP server is running:

# From the main project directory
cd ibmi-mcp-server
npx -y @ibm/ibmi-mcp-server@latest --transport http --tools ./tools

3. Run MCP Client Script

In another terminal, navigate to the client directory (this directory) and run:

uv run mcp_client.py

4. Run the Agent

Configure API Key

Create a .env file in this directory with your OpenAI API key:

# Create .env file
cat > .env << EOF
OPENAI_API_KEY=your-openai-api-key-here
EOF

Get your OpenAI API key:

  1. Visit OpenAI API Keys
  2. Create a new API key
  3. Copy it to your .env file
# Use a custom prompt
uv run agent.py -p "What's my system status?" --model-id "openai:gpt-4o"

# Or run with the default prompt
uv run agent.py

# Get help
uv run agent.py -h

To run the agent with ollama:

uv run agent.py -p "What's my system status?" --model-id "ollama:qwen2.5:latest"

# with gpt-oss
uv run agent.py -p "What's my system status?" --model-id "ollama:gpt-oss:latest"

Available test agents and scripts

  • agent.py: The main example agent that connects to your IBM i MCP server and allows natural language queries.
  • list_tool_annotations.py: A test script that demonstrates how to use tool annotations with the agent.
  • list_toolset_resources.py: A test script that demonstrates how to use toolset resources with the agent.
  • test_auth_agent.py: A test script that demonstrates how to use authentication with the agent.