# Hermes Agent

> Self-improving AI agent that learns from experience and runs anywhere

- Category: Agents & Automation / Assistants & Chatbots
- Language: Python
- Licence: MIT
- Repository: https://github.com/NousResearch/hermes-agent
- Website: https://hermes-agent.nousresearch.com
- Canonical page: https://ai-tldr.dev/tools/hermes-agent/

## In short

Hermes Agent is an open-source personal AI agent built by Nous Research. It sits in the Assistants & Chatbots category of the Agents & Automation tools and is open-source, written in Python and licensed MIT. To get started: install on Linux, macOS, WSL2, or Termux.

## Overview

Hermes Agent is an open-source personal AI agent built by Nous Research. Its standout idea is a built-in learning loop: the agent creates new skills from experience, improves those skills while it uses them, remembers facts you care about, and searches its own past conversations so it gets more useful the longer you work with it.

It is not tied to one machine. You can run Hermes on a cheap VPS, a GPU cluster, or serverless infrastructure that costs almost nothing while idle, then talk to it from the terminal or from messaging apps like Telegram and Discord. You can plug in almost any model provider, from Nous Portal and OpenRouter to OpenAI or your own endpoint, and switch with a single command without changing any code.

## What it does

- Closed learning loop: agent-curated memory, autonomous skill creation after complex tasks, and self-improving skills, with full-text session search and cross-session recall
- Lives where you do: a single gateway process connects Telegram, Discord, Slack, WhatsApp, Signal, and CLI, with voice memo transcription and cross-platform conversation continuity
- Full terminal UI with multiline editing, slash-command autocomplete, conversation history, interrupt-and-redirect, and streaming tool output
- Provider-agnostic models: use Nous Portal, OpenRouter, OpenAI, Hugging Face, or your own endpoint, and switch with `hermes model` with no lock-in
- Built-in cron scheduler for unattended automations like daily reports, nightly backups, and weekly audits, described in natural language
- Runs across six terminal backends (local, Docker, SSH, Singularity, Modal, Daytona), with serverless persistence that hibernates when idle

## Getting started

Hermes ships a one-line installer that sets up Python, Node.js, and the other dependencies it needs, then you start chatting from the terminal. The installer supports Linux, macOS, WSL2, Termux, and native Windows.

1. Install on Linux, macOS, WSL2, or Termux

   Run the install script. It handles uv, Python, Node.js, and the rest of the toolchain for you.

```bash
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
```

2. Install on native Windows

   On Windows, run the PowerShell one-liner. It installs everything, including a portable Git Bash, with no admin rights required.

```powershell
iex (irm https://hermes-agent.nousresearch.com/install.ps1)
```

3. Reload your shell and start chatting

   After installing, reload your shell so the `hermes` command is available, then launch the interactive CLI to start a conversation.

```bash
source ~/.bashrc
hermes
```

4. Pick a model and add messaging

   Choose your provider and model, then start the gateway if you want to talk to Hermes from Telegram, Discord, or other messaging apps.

```bash
hermes model
hermes gateway
```

## When to use it

- Run a personal assistant that you message from Telegram, Discord, or Slack while it works on a cloud VM in the background
- Let the agent build and refine its own reusable skills as you work, so repeat tasks get faster and more reliable over time
- Schedule unattended automations such as daily reports or nightly backups, described in plain language and delivered to any platform
- Host an always-available agent cheaply on a $5 VPS or serverless backend that hibernates when idle and wakes on demand