Mythos — Multi-Agent Framework

A Python framework for building dynamic, memory-aware multi-agent systems with formal orchestration.

Core Concepts

• Agent: A role-based reasoning unit with memory, tools, and communication capabilities
• Pantheon: A collection of agents with defined relationships and workflows
• Oracle: The orchestration engine that routes tasks, manages state, and adapts topology
• Memory: MIRIX-inspired 4-component memory (Core, Episodic, Semantic, Procedural)
• Tool: MCP-native tool abstraction

Quick Start

from mythos import Agent, Pantheon, Oracle

researcher = Agent(name="Athena", role="Researcher", goal="Find accurate information")
writer = Agent(name="Homer", role="Writer", goal="Create compelling narratives")

pantheon = Pantheon(agents=[researcher, writer])
oracle = Oracle(pantheon=pantheon, process="sequential")

result = oracle.run("Write a story about ancient Greece")

Generated by ML Intern

This model repository was generated by ML Intern, an agent for machine learning research and development on the Hugging Face Hub.

• Try ML Intern: https://smolagents-ml-intern.hf.space
• Source code: https://github.com/huggingface/ml-intern

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "huronvalley21/mythos"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)

For non-causal architectures, replace AutoModelForCausalLM with the appropriate AutoModel class.