machine-core

A flexible agent framework for building AI agents with MCP (Model Context Protocol) integration.

Features

• 🎯 Clean Architecture - Separation of infrastructure (AgentCore) and execution patterns (BaseAgent)
• 🔧 Flexible Configuration - Environment variables, direct parameters, or runtime overrides
• 🔌 MCP Integration - Easy integration with MCP servers and tools
• 🚀 Multiple Agent Types - Chat, CLI, Receipt Processor, Twitter Bot, Memory Master, RAG Chat
• 📦 Reusable Package - Install once, use in multiple projects

Installation

# From the machine-core directory
uv add git+https://github.com/samletnorge/machine-core.git

# Or with
uv sync

Quick Start

Basic Usage (Environment Config)

from machine_core.agents import ChatAgent

# Loads config from environment variables
agent = ChatAgent()

# Run streaming query
async for event in agent.run("What is quantum computing?"):
    if event['type'] == 'text_delta':
        print(event['content'], end='', flush=True)

Custom Configuration

from machine_core import AgentConfig
from machine_core.agents import CLIAgent

# Create custom config
config = AgentConfig(
    max_iterations=20,
    timeout=3600.0,
    max_tool_retries=10
)

# Pass to agent
agent = CLIAgent(
    model_name="llama3.2:latest",
    agent_config=config
)

result = await agent.run("Analyze this data")

Configuration

Machine Core supports three configuration methods:

1. Environment Variables (.env file)
2. Direct Parameters (runtime override)
3. Partial Overrides (mix and match)

See CONFIGURATION.md for detailed examples.

Environment Variables

# Agent Config
AGENT_MAX_ITERATIONS=10
AGENT_TIMEOUT=604800.0
AGENT_MAX_TOOL_RETRIES=15
AGENT_ALLOW_SAMPLING=true

# LLM Config
LLM_PROVIDER=ollama
LLM_MODEL=gpt-oss:latest

# Embedding Config
EMBEDDING_PROVIDER=ollama
EMBEDDING_MODEL=nomic-embed-text

Available Agents

Agent	Description	Use Case
ChatAgent	Streaming chat	Streamlit UI, web chat
CLIAgent	Non-streaming	Terminal, cron jobs
ReceiptProcessorAgent	Vision + queue	Document analysis
TwitterBotAgent	Scheduled posting	Social media automation
RAGChatAgent	Knowledge graph	Q&A, support
MemoryMasterAgent	Knowledge extraction	Graph maintenance

Creating Custom Agents

from machine_core.core.agent_base import BaseAgent

class MyCustomAgent(BaseAgent):
    def __init__(self):
        super().__init__(
            system_prompt="Your custom prompt",
            mcp_config_path="mcp_custom.json"
        )
    
    async def run(self, task: str):
        """Your custom execution logic."""
        result = await self.run_query(task)
        return result

Architecture

AgentCore (infrastructure)
  ├─ MCP toolset loading/validation
  ├─ Model/provider configuration
  ├─ Embedding backend setup
  └─ Agent instance creation

BaseAgent (execution patterns)
  ├─ run() [abstract - implement per agent]
  ├─ run_query() [sync execution]
  ├─ run_query_stream() [streaming execution]
  └─ Helper methods

ConcreteAgent (implementations)
  └─ Implements run() using base patterns

License

LICENSE