Repolode
Submit
NousResearch

NousResearch/hermes-agent

A self-improving AI agent that creates and refines skills from experience, runs on many model providers, and works through CLI and messaging apps.

Open-source alternative to:ChatGPT

View on GitHubWebsite
Stars
255,284
Forks
55,575
Open issues
50,096
Last commit
5 hours ago

What is it?

Hermes Agent is an AI agent built by Nous Research that keeps learning across sessions. It creates skills from completed tasks, improves them during use, saves knowledge to persistent memory, and searches past conversations to recall earlier context.

You can switch LLM providers with a single command, including Nous Portal, OpenRouter, OpenAI, or a custom endpoint. It offers a terminal interface with multiline editing and streaming tool output, and a gateway process that connects the same agent to Telegram, Discord, Slack, WhatsApp, Signal, and the CLI. It also includes a cron scheduler for unattended tasks and can run subagents in parallel.

Execution is not limited to a laptop. The README lists seven terminal backends: local, Docker, SSH, Singularity, Modal, Daytona, and Vercel Sandbox, with Daytona and Modal offering serverless environments that sleep when idle.

Key features

  • 01Creates reusable skills from completed tasks and refines them during use
  • 02Persistent memory plus full-text search over past sessions
  • 03Switches between many LLM providers with a single command
  • 04Terminal UI and messaging gateway for Telegram, Discord, Slack, WhatsApp, and Signal
  • 05Built-in cron scheduler with delivery to chat platforms
  • 06Connects to MCP servers for additional tools

Use cases

  • Running a personal assistant that you reach from Telegram while the agent works on a cloud VM
  • Automating recurring reports, backups, or audits described in natural language
  • Delegating parallel research or coding subtasks to isolated subagents
  • Testing agent workflows across different model providers

Quick start

  1. 1

    Install with the Linux/macOS/WSL2 one-liner

    curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
  2. 2

    Reload your shell

    source ~/.bashrc # reload shell (or: source ~/.zshrc)
  3. 3

    Start chatting with the agent

    hermes

Things to know

  • Using most features requires an LLM provider API key or a Nous Portal subscription.
  • Messaging platform use requires setting up a gateway and bot credentials.

Topics

#ai-agent#llm#cli#messaging-bot#skills#mcp#automation#self-hosted#ai#ai-agents

FAQ

What is hermes-agent?
A self-improving AI agent that creates and refines skills from experience, runs on many model providers, and works through CLI and messaging apps. Hermes Agent is an AI agent built by Nous Research that keeps learning across sessions. It creates skills from completed tasks, improves them during use, saves knowledge to persistent memory, and searches past conversations to recall earlier context.
Is hermes-agent free to use? What license does it have?
hermes-agent is open source under the MIT license. Check the license terms before using it in commercial products.
Is hermes-agent still maintained?
Actively maintained. The last commit to hermes-agent was on Oct 20, 2026.
What can I use hermes-agent for?
Running a personal assistant that you reach from Telegram while the agent works on a cloud VM Automating recurring reports, backups, or audits described in natural language Delegating parallel research or coding subtasks to isolated subagents Testing agent workflows across different model providers
How do I install hermes-agent?
Install with the Linux/macOS/WSL2 one-liner: curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash Reload your shell: source ~/.bashrc # reload shell (or: source ~/.zshrc) Start chatting with the agent: hermes
Where can I get hermes-agent?
The source code and installation instructions are on GitHub: https://github.com/NousResearch/hermes-agent

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects