What is Hermes Agent?

Hermes Agent is an open-source autonomous AI agent developed by Nous Research. Unlike traditional coding copilots or simple chatbots, it is designed as a persistent, self-evolving companion that runs on your server (or VPS), learns from interactions, creates its own skills, and continuously improves over time.

It combines long-term memory, procedural skill creation, and cross-session recall to become more useful the longer you use it.

Key Features

Built-in Learning Loop: Automatically creates new skills after completing complex tasks and refines them during use.
Long-Term Memory: Uses FTS5 session search with LLM summarization for intelligent recall across conversations.
Persistent & Server-Based: Runs on Linux, macOS, or WSL; accessible via messaging platforms like Telegram, Discord, Slack, and WhatsApp.
Skill System: Browse, invoke, and manage skills with commands like /skills or /<skill-name>.
Flexible LLM Support: Works with OpenAI, Anthropic (Claude), OpenRouter, or self-hosted models (Ollama, vLLM, etc.).
Autonomous Operation: Can work in the background on a cloud VM while you interact from mobile apps.
Interruptible: Easily stop current tasks with Ctrl+C or new messages.

How It Works

Hermes Agent maintains a deepening model of you across sessions. After finishing tasks, it can autonomously generate skills that encapsulate the learned process. These skills self-improve through usage and can be invoked directly. Its memory system allows it to search and summarize past interactions intelligently.

Use Cases

Personal automation and research assistant that improves daily
Long-running background tasks on a cheap VPS
Building a personal knowledge base through conversational learning
Multi-platform access for on-the-go interaction with your agent
Experimenting with self-improving agent architectures

Installation

Quick install on Linux/macOS/WSL:

curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash

Supports deployment on low-cost servers or GPU clusters. Runs efficiently when idle.

Why Hermes Agent?

In a world of disposable chatbots, Hermes Agent stands out by treating the agent as a growing entity rather than a stateless tool. It is MIT licensed and fully open source, encouraging community contributions and customization.

Whether you're a developer, researcher, or power user looking for an AI that truly "grows with you," Hermes Agent offers a compelling vision for the next generation of personal AI agents.