✓ You're in — here's the full guide. Check your inbox for The Input Daily.

Free guide

Hermes Agent: the open-source AI agent that gets better every time you use it

Most AI tools forget everything the second you close the chat. Hermes doesn't — it remembers, and it compounds.

By Adan · adanxai · Free, given away in full

Built by Nous Research, Hermes is an open-source AI agent that uses real tools, remembers previous work, and improves over time by turning successful workflows into reusable Skills. Instead of acting like another chatbot, it behaves more like a junior developer that learns from experience.

210k+ GitHub stars
MIT fully open source
Feb 2026 released

That star count makes it one of the fastest-growing open-source AI projects ever — for context, it's already ahead of VS Code, which took a decade to get there.

Installation

Getting started takes less than a minute. Run:

curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash

The installer configures everything automatically — no config files to edit before you begin. It supports Linux, macOS, WSL2, and Termux, and there's a native Windows install as well.

Quick note: that command pipes a script straight from the internet into your shell. It's the official installer from Nous Research's own domain, so it's fine — but that's exactly the habit worth being careful with generally. If you'd rather read before you run, open the URL in a browser first.

Choosing your model

One of Hermes' biggest strengths: it isn't tied to a single provider. You can switch between:

  • OpenAI
  • Anthropic
  • Google Gemini
  • DeepSeek
  • OpenRouter
  • Local LLMs

Changing models doesn't mean rebuilding your workflow, so you can always use whichever model is best for the task in front of you.

Your first conversation

Once installed, run:

hermes

From there you can ask it to:

  • Read entire repositories
  • Debug code
  • Search the web
  • Edit files
  • Summarize folders
  • Execute terminal commands
  • Build projects

Unlike a traditional chatbot, Hermes actually performs these actions with real tools instead of describing what it would do.

Persistent memory

Hermes remembers your previous work. Instead of starting from scratch every session, it keeps context on your projects, your coding style, and past conversations. The more you work with it, the more personalized it gets.

Skills: the secret weapon

Every time Hermes solves a problem, it can turn that solution into a reusable Skill — think of them as AI-powered workflows you can run again later:

  • /debug
  • /github-pr
  • /deploy
  • /k8s

It can also download Skills built by the community, improve existing ones, and merge similar Skills automatically. Over time your agent gets more capable because it's building its own library of experience.

Beyond the terminal

Hermes doesn't have to live in your terminal. Run the gateway and talk to it from:

  • Discord
  • Slack
  • Telegram
  • WhatsApp
  • Signal
  • Email

Same persistent agent, same shared memory, from anywhere.

Pro tips

  • Run hermes doctor when something breaks — it diagnoses common issues automatically.
  • Use Docker as your terminal backend for safer, sandboxed execution.
  • Enable /voice for hands-free conversations.
  • Host on Modal or Daytona to keep costs low while idle.
  • Use hermes --continue to resume a previous session without losing context.

Why it stands out

Most AI assistants reset after every conversation. Hermes compounds. Every task and every solved problem makes it more useful than it was yesterday. Instead of rewriting the same prompts forever, you're building an assistant that actually understands your workflow.

If you want an open-source agent that feels less like a chatbot and more like a teammate, this is one of the most impressive projects out there right now.

Want the next one first?

The Input Daily — a 5-minute AI brief every weekday. 3 stories that matter, 1 tool worth trying, 1 prompt to steal.

Free. Every morning. Unsubscribe in one click. Privacy Policy.