Free, open-source AI agent from Nous Research with persistent memory, auto-generated skills, and multi-platform access.
Hermes Agent is a free, open-source AI agent from Nous Research that runs on macOS 12+, Windows 10/11, and Linux. It keeps a persistent memory of your projects, generates skills from work it has already done, and reaches you through Telegram, Discord, Slack, WhatsApp, Signal, email, and a command-line interface.
Natural-language scheduling handles reports, backups, and briefings that run unattended through a gateway. Work can be split across isolated subagents with their own conversations, terminals, and Python RPC scripts. Sandboxing spans five backends, and the agent also browses the web and handles vision, image generation, and text-to-speech.
The download page lists macOS 12+, Windows 10/11 (including an .exe download), and Linux for any distro.
Yes. Nous Research describes it as a free, open-source AI agent, and the source is available on GitHub.
The listed surfaces are Telegram, Discord, Slack, WhatsApp, Signal, email, and a command-line interface, with the site noting a growing list of platforms.
The agent learns your projects and auto-generates skills, and according to the site it does not forget how it solved a problem.
Scheduling is written in natural language and covers things like reports, backups, and briefings, which run unattended through the gateway.
Subagents run isolated with their own conversations, terminals, and Python RPC scripts, which the site describes as supporting zero-context-cost pipelines.
There are five: local, Docker, SSH, Singularity, and Modal, with container hardening and namespace isolation.
The listed capabilities include web search, browser automation, vision, image generation, text-to-speech, and multi-model reasoning.
Support runs through the Nous Research Discord community and the GitHub repository, where the source code is hosted.
