← Back to glossary
+Suggest a term
Tool·Agents & Automation·Added today

Hermes Agent

Also known as: Nous Hermes Agent, Hermes Agent Framework

An open-source autonomous AI agent framework from Nous Research, built around persistent memory that compounds over time. Unlike stateless single-session agents, Hermes Agent is designed to grow more effective at long-term engineering tasks the longer it runs.

Hermes Agent was released by Nous Research in February 2026 and crossed 105,000 GitHub stars by May 2026, making it one of the most-starred open-source agent frameworks. Its central design idea is the 'agent that grows with you': rather than starting fresh each session, Hermes maintains a persistent context and memory layer that accumulates knowledge about your codebase, your preferences, and your project history across sessions.

This persistent memory architecture is the key differentiator from stateless tools like Claude Code or standard Cline setups. In practice it means the agent gets progressively better at your specific repo over time — it learns which patterns you prefer, which components are connected, and which prior decisions to avoid revisiting. For long-running engineering projects, that compounding context reduces the need to re-explain your setup on every session.

Hermes Agent is built on top of Nous Research's open-weight Hermes model family and is designed to run as a local or self-hosted agent, though it can be pointed at other model backends. As a specialist tool, it requires more setup than commercial alternatives and is primarily used by developers comfortable with running their own agent infrastructure. The framework is particularly popular in the open-source AI community as a reference implementation for persistent-memory agent design.

This definition is AI-generated and refreshed weekly. It may contain inaccuracies. Use your own judgment, especially for production decisions.
Related terms
Agent frameworkAgent memoryAgent stateLong-running agentOpen-source model