Modular AI command system with expert autonomous agents powered by Mistral API.
Neuronet is a scalable and modular Central AI System built in Python, powered by Mistral APIs. It utilizes a hierarchy of intelligent autonomous agents, each specialized in a particular domain, all coordinated by a central controller.
Neuronet is designed to decompose complex user objectives into manageable subtasks and intelligently distribute them to the appropriate domain-specific agents. Each agent operates independently while communicating with the central controller and, when needed, with other agents.
- Central Controller: Receives user instructions and determines strategy.
- Specialized Agents: Modular units acting as domain experts (e.g., research, coding, evaluation).
- Agent Communication Layer: Enables message passing and coordination.
- Task Management Engine: Breaks down objectives, routes tasks, tracks progress.
- Python 3.11+
- Mistral API
- LangChain (planned)
- Vector Stores (e.g., FAISS/ChromaDB)
- AsyncIO for concurrent task handling
- Strategy Agent: Plans subtask execution flow.
- Research Agent: Gathers and summarizes information.
- Coding Agent: Writes or refactors code.
- Evaluator Agent: Reviews outputs for quality and completeness.
- Memory Agent: Stores context for long-running tasks.
Each agent is plug-and-play and can be swapped or extended independently.
neuronet/ │ ├── core/ # Central controller and task manager ├── agents/ # Modular agents ├── utils/ # Helper functions/utilities ├── configs/ # API keys, constants, and parameters ├── examples/ # Sample objective runs └── README.md
Neuronet aims to be a base for building next-generation AI systems capable of handling complex objectives autonomously through a team of collaborative, intelligent agents.
Open to collaborations! Feel free to fork the repo and raise PRs or suggestions.
MIT License