Overview

Hermes Agent is an open‑source AI agent developed by Nous Research. It can be run locally or in the cloud and is available for macOS, Windows, and Linux. It is licensed under MIT, so you can download, modify, and deploy it at no cost.

The agent is designed to keep a persistent memory of your projects. When you start a conversation it stores the dialogue, any generated skills, and the context it used to solve problems. Next time you return, the agent remembers what it did before and can rebuild the same workflow without re‑input. It can also automatically create new skills based on the tasks you perform, reducing the need to write custom code for every new use case.

You interact with Hermes through a desktop client, or directly from a terminal, and can connect it to a variety of messaging platforms. Supported gateways include Telegram, Discord, Slack, WhatsApp, Signal, email, and a command‑line interface. The integration docs walk you through adding each gateway so you can chat with the agent from the surface that suits you best.

Beyond conversation, Hermes offers a set of automation features. It can schedule tasks using natural‑language commands, such as “run a backup every Sunday at 2 am” or “generate a weekly report and email it.” The agent can spawn isolated sub‑agents with their own terminal sessions and Python RPC scripts, allowing you to build pipelines that run without context overhead. Web search, browser automation, vision, and image generation are also available, giving you a multi‑model reasoning stack that can browse the internet or interact with a browser as needed.

Hermes can run on several backends: locally, in Docker, over SSH, using Singularity, or on Modal. Each backend provides container hardening and namespace isolation, keeping experiments separate from one another. When you stop a local instance its data remains on disk, so you can restart and pick up where you left off.

Key features

Persistent memory of conversations, skills, and problem solutions
Self‑created skills that grow with your usage
Messaging gateway for Telegram, Discord, Slack, WhatsApp, Signal, email, and CLI
Natural‑language scheduling for reports, backups, and briefings
Isolated sub‑agents with separate terminals and Python RPC
Web search, browser automation, vision, image generation, and text‑to‑speech
Multiple backend options: local, Docker, SSH, Singularity, Modal