MedhaAI is a flexible and powerful multi-agent AI framework for building advanced AI applications. It provides a set of tools and abstractions that allow developers to create complex, collaborative AI systems with ease.
- Multi-agent system with prioritized agents
- Support for multiple LLM providers (OpenAI, Anthropic)
- Flexible memory systems
- Advanced planning and execution modules
- Customizable toolkit system with advanced web scraping capabilities
- Live streaming of agent activities
- Performance monitoring and tracing
- Asynchronous operations for improved performance
To install MedhaAI, simply use pip:
pip install medhaaiHere's a simple example to get you started with MedhaAI:
import asyncio
from medhaai import MedhaConfig, BaseLLM, BaseAgent, MedhaMultiAgentSystem
async def main():
config = MedhaConfig()
llm = BaseLLM.create("openai", api_key=config.openai_api_key)
agent = BaseAgent.create("general", name="GeneralAgent", role="assistant", llm=llm)
mas = MedhaMultiAgentSystem(task_decomposition_llm=llm)
mas.add_agent(agent)
result = await mas.run_task("Research and summarize the latest advancements in quantum computing")
print(result)
asyncio.run(main())For more detailed information and advanced usage, please refer to our documentation.
Contributions are welcome! Please see our Contributing Guide for more details.
This project is licensed under the MIT License - see the LICENSE file for details.