mz.
I have been using Hermes Agent by Nous Research for a while now, and it is hands down the most capable AI agent I have worked with. It runs in the terminal, connects to Telegram, and has persistent memory, tools, cron jobs, and skills. It is fully open-source with 206k+ GitHub stars. Here is how to install it.
GitHub: github.com/NousResearch/hermes-agent
Official site: https://hermes-agent.nousresearch.com
Docs: https://hermes-agent.nousresearch.com/docs
What Is Hermes Agent?
Hermes Agent is a standalone AI assistant by Nous Research that runs locally on your machine. It is not a chatbot in a browser window. It lives in your terminal, has access to your filesystem, can execute code, browse the web, send messages, and run scheduled tasks. Think of it as an AI co-pilot that actually does things instead of just talking about them.
Key features:
- Terminal + Desktop App – Runs in CLI or as a native macOS/Windows/Linux app
- Messaging Platforms – Connect Telegram, Discord, WhatsApp, Signal, and more
- Persistent Memory – Remembers facts and preferences across sessions
- Skills System – Install and write custom skills for recurring tasks
- Cron Jobs – Schedule automated tasks like daily briefings
- Tool Gateway – Web search, image generation, TTS, cloud browser
- MCP Support – Connect to any MCP server for extended tooling
- 200+ Models – Works with OpenAI, Anthropic, DeepSeek, and more
Quick Install (Linux / macOS)
One curl command and you are done. The installer handles everything – Python, Node.js, dependencies, and the global hermes command.
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bashOnce it finishes, reload your shell and start chatting:
source ~/.bashrc
hermesWindows Install
Open PowerShell as Administrator and run:
iex (irm https://hermes-agent.nousresearch.com/install.ps1)Alternatively, download the Hermes Desktop installer from the official site for a GUI-based setup on macOS, Windows, or Linux.
Post-Install: Choose a Model
After installation, you need to pick an LLM provider. The easiest path is the Nous Portal which gives you 300+ models with one subscription:
hermes setup --portalThis logs you in via OAuth, sets Nous as your provider, and enables the Tool Gateway (web search, image generation, TTS, cloud browser) in one command.
If you prefer to bring your own API keys, run the setup wizard:
hermes setupConfiguration Commands
Hermes has dedicated commands for every setting:
| Command | What It Does |
|---|---|
hermes model | Choose your LLM provider and model interactively |
hermes tools | Enable or disable specific tools |
hermes gateway setup | Connect messaging platforms (Telegram, Discord, etc.) |
hermes config set | Set individual config values |
hermes portal info | Check what services are wired up |
Install Layout
For a normal per-user install, the layout looks like this:
| Path | Purpose |
|---|---|
~/.hermes/hermes-agent/ | Code and dependencies |
~/.local/bin/hermes | Global hermes command (symlink) |
~/.hermes/ | Config, skills, sessions, cron jobs |
If you run the installer as root (sudo curl ... | sudo bash), it uses an FHS layout under /usr/local/lib/hermes-agent/ with the binary at /usr/local/bin/hermes. User config still lives under each user's ~/.hermes/.
Updating Hermes Agent
Keeping it updated is simple:
hermes updateThis pulls the latest code from GitHub, updates dependencies, and preserves all your config, skills, and session data. No manual git work needed.
Why I Like It
- One command install – Everything auto-detected and set up
- Runs in my terminal – No browser tab needed, stays out of my way
- Telegram integration – I chat with it from my phone
- Cron jobs – Daily CVE briefings delivered automatically
- Open source – MIT license, active community, constant updates
Honestly, Hermes Agent changed how I work. Having an AI that can execute code, browse the web, send messages, and remember what I told it last week is a game changer. If you spend time in a terminal, give it a shot.