Overview
Nous Research is a leader in the American open source AI movement. They train world-class open source language models and build infrastructure to coordinate distributed, unbiased training.
Their mission: advance human rights and freedoms by creating and proliferating open source language models, supporting their unrestricted availability and use.
Key Projects
Hermes Agent (95k stars)
“The agent that grows with you” — A general-purpose AI agent with skill-based architecture.
- 95,592 GitHub stars
- MIT licensed
- Supports skill loading, MCP tools, multi-turn conversations
- Built-in environments and plugins
DisTrO
Distributed Training Over-The-Internet — Training language models across distributed nodes.
Atropos
Language Model Reinforcement Learning Environments framework for collecting and evaluating LLM trajectories through diverse environments.
Autoreason
Autoresearch for subjective domains — exploring reasoning in ambiguous contexts.
Hermes-Agent-Self-Evolution
Evolutionary self-improvement for Hermes Agent — optimize skills, prompts, and code using DSPy + GEPA.
Research Focus
- Model architecture: Novel transformer designs and efficiency improvements
- Data synthesis: Generating high-quality training data
- Fine-tuning: Adapting models to specific domains
- Reasoning: Chain-of-thought, tool use, multi-step problem solving