logicore 0.1.1

A modular, multi-provider AI agent framework for building intelligent tool-using agents with Gemini, Groq, Ollama, Azure OpenAI, and Anthropic.

🧠 Logicore

A modular, multi-provider AI agent framework for Python

Build intelligent, tool-using AI agents that work across Gemini, Groq, Ollama, Azure OpenAI, and Anthropic — with a single unified API.

📖 Documentation · 🐛 Report Bug · 💡 Request Feature

---

✨ Why Logicore?

Logicore is a lightweight, production-ready Python framework for building AI agents that can use tools, remember context, and work with any major LLM provider. No vendor lock-in — swap providers with one line.

🔑 Key Features

Feature	Description
Multi-Provider	Gemini, Groq, Ollama, Azure OpenAI, Anthropic — one interface
Tool Use	Built-in file, web, git, PDF, Office, and code execution tools
MCP Support	Model Context Protocol for dynamic tool discovery
Streaming	Real-time token streaming with async callbacks
Vision	Multimodal image understanding across supported models
Memory	Session persistence and simple memory stores
Telemetry	Built-in execution tracing and walkthrough generation
Skills	Modular, reusable skill packs for domain-specific tasks
Hot Reload	Live code reloading during development

---

🚀 Quick Start

Installation

# Core framework
pip install logicore

# With a specific provider
pip install logicore[gemini]    # Google Gemini
pip install logicore[groq]      # Groq
pip install logicore[ollama]    # Ollama (local)
pip install logicore[azure]     # Azure OpenAI / Anthropic

# Everything
pip install logicore[all]

Your First Agent

from logicore import BasicAgent, create_agent, tool

# Define a custom tool
@tool
def get_weather(city: str) -> str:
    """Get current weather for a city."""
    return f"It's sunny and 24°C in {city}!"

# Create an agent with Groq
agent = create_agent(
    provider="groq",
    model="llama-3.3-70b-versatile",
    tools=[get_weather]
)

# Chat with your agent
response = await agent.chat("What's the weather in Tokyo?")
print(response)

---

Multi-Provider Flexibility

from logicore import Agent, SmartAgent
from logicore import GroqProvider, GeminiProvider, OllamaProvider

# Swap providers without changing your agent code
groq = GroqProvider(model_name="llama-3.3-70b-versatile")
gemini = GeminiProvider(model_name="gemini-2.0-flash")
ollama = OllamaProvider(model_name="llama3.2")

agent = Agent(provider=groq)  # or gemini, or ollama
response = await agent.chat("Explain quantum computing")

MCP Integration

from logicore import MCPAgent, GroqProvider

# Agent with Model Context Protocol servers
agent = MCPAgent(
    provider=GroqProvider(model_name="llama-3.3-70b-versatile"),
    mcp_config="mcp.json"
)
response = await agent.chat("Search for recent AI papers")

🛠️ Built-in Tools

Logicore ships with a comprehensive set of ready-to-use tools:

• File System — Read, write, list, search files
• Web — Search the web, fetch URLs, image search
• Git — Run git commands programmatically
• Code Execution — Execute Python and shell commands safely
• Documents — Read PDFs, DOCX, XLSX, and more
• PDF Tools — Merge and split PDF files
• Office Tools — Create and edit Word/Excel documents

---

🏗️ Agent Types

Agent	Best For
BasicAgent	Simple single-turn tool-calling agents
Agent	Full-featured agents with memory and tools
SmartAgent	Autonomous multi-step reasoning agents
CopilotAgent	Interactive copilot with step-by-step execution
MCPAgent	Agents with dynamic MCP tool discovery

---

📖 Documentation

Full documentation available at: https://rudramodi360.github.io/Agentry/

---

🤝 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.

---

Built with ❤️ by RudraModi360

⭐ Star this repo if you find it useful!