Brethof AI Reviews

Hermes Agent CLI review

Open, any model, and it learns as it works

Not sponsoredAI Agents · Developer Tools

Tested on Linux · Free · Open source

Hermes Agent is Nous Research's open-source AI agent for the terminal. It runs on almost any model, writes its own skills, keeps memory in plain files, and works from chat apps like Telegram. We tested it on Linux with an open model through Ollama.

The good

  • Any model, open source. We like that Hermes Agent is MIT licensed and works with over 40 providers, including our own Ollama endpoint, so the model is entirely your choice.
  • Talk to it from your chat apps. One gateway connects Hermes Agent to Telegram, Discord, Slack, WhatsApp and more, and scheduled jobs deliver results straight into those chats.
  • It learns in the open. We like that it writes its own reusable skills, keeps memory in plain Markdown files you can read and edit, and installs curated plugins pinned to a commit.

The not so good

  • Linux desktop: build it yourself. One real downside: Hermes Desktop has no ready-made Linux installer. On Linux you build it yourself with one command, hermes desktop, while macOS and Windows get an installer.
  • Big and fast-moving. Hermes Agent moves fast: five releases in eight days, and the latest notes say the one before was built from the wrong commit. With over sixty commands and a config over two thousand lines, it's a lot to learn.

Our verdict

Recommended

Free, MIT licensed, any model including local ones, and it works from more than twenty chat apps, with scheduled jobs. Downsides: on Linux you build the desktop app yourself, and it is big and changes every few days.

Reviewed: Hermes Agent CLI 0.22.1+83 (main, db44c46) · facts checked 2026-10-15

← All reviews

Hear it when it ships

New releases, real benchmarks and the occasional deep-dive. No spam, unsubscribe in one click.

Everything we build

External:   YouTube · GitHub