# Hermes Agent

> The self-improving AI agent built by Nous Research. The only autonomous agent with a built-in learning loop: it creates skills from experience, improves them during use, and builds a deepening model of who you are across sessions.

Hermes Agent is open source (MIT), written in Python, and runs on Linux / macOS / WSL2. It is the work of Nous Research (nousresearch.com) — the lab behind the Hermes, Nomos, and Psyche model families. Hermes Agent is distinct from those LLMs: it is an *agent framework* that can be driven by any OpenAI-compatible LLM provider (Claude, GPT, Gemini, Qwen, DeepSeek, etc.). Requires a ≥64k context-window model.

This site, hermes-ai.net, is a community-maintained multilingual landing and docs site. It is not affiliated with Nous Research and is not the official Hermes Agent release domain. Official English documentation lives at https://hermes-agent.nousresearch.com/docs — we translate and supplement it for Chinese, Japanese, Korean, Spanish, French, German, and Russian readers.

## Core concepts

- [Self-improving learning loop](https://hermes-ai.net/docs/features): Agent-curated memory with periodic nudges, autonomous skill creation, skill self-improvement during use, FTS5 cross-session recall, Honcho dialectic user modeling.
- [Skills system](https://hermes-ai.net/docs/features): Procedural memory the agent creates and reuses. Compatible with agentskills.io — skills are portable, shareable, community-contributed Markdown files.
- Messaging gateway: One gateway delivers to 15+ platforms — Telegram, Discord, Slack, WhatsApp, Signal, Matrix, Mattermost, Email, SMS, DingTalk, Feishu, WeCom, BlueBubbles, Home Assistant.
- MCP integration: Attach any Model Context Protocol server to extend Hermes's toolset. Allowlist filtering keeps the tool surface small.
- Runtime backends: Six execution backends — local, Docker, SSH, Daytona, Singularity, Modal. Serverless backends hibernate when idle.
- Scheduled automations: Built-in cron system with delivery to any messaging channel.

## Installation (canonical)

```
curl -fsSLO https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh
less install.sh
bash install.sh
```

Then `hermes setup` for the interactive wizard, or `hermes model` / `hermes tools` / `hermes env` to configure parts individually. Launch with `hermes` (classic CLI) or `hermes --tui` (terminal UI). Resume a previous session with `hermes --continue`. Alternative installs: `pipx install hermes-agent`, or the `nousresearch/hermes-agent` Docker image.

## Docs

- [Quickstart (English)](https://hermes-ai.net/docs/quickstart): 60-second path from zero to first conversation.
- [Installation](https://hermes-ai.net/docs/installation): One-liner, pipx, Docker, and the six terminal backends.
- [Features](https://hermes-ai.net/docs/features): Learning loop, skills, messaging, MCP, cron, runtime backends.
- [FAQ](https://hermes-ai.net/docs/faq): What Hermes is, how it differs from copilots, supported providers, self-hosting, licensing.

## Translations

Same four docs pages in: [中文](https://hermes-ai.net/zh/docs/quickstart) · [日本語](https://hermes-ai.net/ja/docs/quickstart) · [한국어](https://hermes-ai.net/ko/docs/quickstart) · [Español](https://hermes-ai.net/es/docs/quickstart) · [Français](https://hermes-ai.net/fr/docs/quickstart) · [Deutsch](https://hermes-ai.net/de/docs/quickstart) · [Русский](https://hermes-ai.net/ru/docs/quickstart).

## Authoritative sources

- Source code and issues: https://github.com/NousResearch/hermes-agent
- Official desktop download page: https://hermes-agent.nousresearch.com/desktop
- Official docs (English, canonical): https://hermes-agent.nousresearch.com/docs
- Nous Research: https://nousresearch.com
- Community Discord: https://discord.gg/NousResearch
- Skills registry: https://agentskills.io

## Licensing

MIT. Free to self-host, fork, and redistribute. You pay only for LLM API calls made with your own provider keys.