Hermes Agent is an open-source, self-hosted AI agent created by Nous Research, the lab behind the Hermes, Nomos, and Psyche model families. Designed to become more useful through experience, it maintains cross-session memory and can create and refine procedural skills as it works. Its tools support web search, browser automation, vision, image generation, text-to-speech, and multi-model reasoning, while parallel sub-agents can handle separate workstreams.

Users can interact with Hermes through Telegram, Discord, Slack, WhatsApp, Signal, email, or a command-line interface, continuing conversations across platforms. It offers terminal backends for local, Docker, SSH, Daytona, Singularity, and Modal environments, alongside features such as container hardening, namespace isolation, and command approval workflows. The project is MIT-licensed, self-hosted, and states that it collects no telemetry. Installation is offered for Linux, macOS, and WSL2. The site also provides community guides, documentation links, a skills hub, release updates, comparisons with other agent frameworks, and information about the broader ecosystem. It describes itself as an independent community project, not affiliated with Nous Research.