Devtoolwatch BoardGet it every Monday

Hermes Agent review

Hermes Agent is an open-source, self-hosted AI agent by Nous Research with persistent memory, self-created skills, and a messaging gateway.

Updated Nov 1, 2026. Review written Oct 22, 2026 from the project's own sources; usage data refreshed every Monday.

TL;DR

Hermes Agent is a self-hosted AI agent from Nous Research that runs in a terminal UI, a desktop app, or messaging apps like Telegram and Slack. It is for developers who want an agent that runs on servers or cloud sandboxes. Hermes Agent creates and improves skills from experience and keeps memory across sessions.

Hermes Agent at a glance

Category
Coding agents
Maintained by
Nous Research (organization, @NousResearch on GitHub)
License
MIT
Language
Python, TypeScript, JavaScript
Repository
github.com/NousResearch/hermes-agent
Website
hermes-agent.nousresearch.com
PyPI package
hermes-agent 0.19.0
Latest release
v0.22.4 on GitHub, Oct 21, 2026
First published
Jul 22, 2025

What Hermes Agent does

  • Hermes Agent creates skills after complex tasks, improves them during use, and is compatible with the agentskills.io open standard.
  • Hermes Agent keeps agent-curated memory and searches past sessions with FTS5 search and LLM summarization for cross-session recall.
  • Hermes Agent connects to Telegram, Discord, Slack, WhatsApp, Signal, Email, and CLI from a single gateway process.
  • Hermes Agent includes a built-in cron scheduler for natural-language automations delivered to any connected platform.
  • Hermes Agent spawns isolated subagents for parallel work and runs Python scripts that call tools via RPC.
  • Hermes Agent works with many model providers, including Nous Portal, OpenRouter, OpenAI, and custom endpoints, switched with hermes model.
  • Hermes Agent can connect any MCP server for extended capabilities.

How it works

Hermes Agent runs as a terminal UI started with hermes, as Hermes Desktop on macOS and Windows, or as a gateway that messaging apps talk to. Commands execute on one of seven terminal backends: local, Docker, SSH, Singularity, Modal, Daytona, or Vercel Sandbox. Models come from your chosen provider, and Nous Portal can also route web search, image generation, TTS, and a cloud browser.

How to install and start using Hermes Agent

Linux, macOS, WSL2

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

Windows (PowerShell)

iex (irm https://hermes-agent.nousresearch.com/install.ps1)

Requirements

  • An account or API key with a model provider, or a Nous Portal subscription.
  • The PyPI package requires Python 3.11 or later and below 3.14.
  • Hermes Desktop requires macOS 12+ or Windows 10/11; Linux installs via terminal.
  • Android use requires an aarch64 device with Termux and the signed APT repository.

First steps

  1. Reload your shell with source ~/.bashrc, then run hermes to start chatting in the terminal UI.
  2. Run hermes setup to launch the full setup wizard, or hermes setup --portal to use Nous Portal.
  3. Run hermes model to choose your LLM provider and model.
  4. Run hermes gateway setup and hermes gateway start to talk to Hermes Agent from messaging platforms.

Who Hermes Agent is for

  • Developers who want an always-on agent on a VPS or cloud VM, reachable from Telegram or Slack.
  • Users who want to switch between model providers without code changes.
  • Teams scheduling unattended tasks like daily reports, nightly backups, or weekly audits.
  • OpenClaw users, since Hermes Agent can import settings, memories, skills, and API keys.

Limitations to know

  • Hermes Agent is free, but model providers and optional hosted services have their own pricing.
  • Hermes Agent tasks that need an active agent will not run while a local agent is stopped.
  • On Windows, antivirus may flag the bundled uv.exe as a false positive, requiring a folder exclusion.
  • Hermes Agent v0.22.2 has a known issue: sessions run under the 0.22.1 gateway cannot be deleted.

Is Hermes Agent growing? Usage data

In the week of Oct 25–31, 2026, Hermes Agent ranked #227 on the Devtoolwatch Board and #77 in Coding agents, with a momentum score of 40 out of 100. Its main package had 35.7k downloads that week, −59% against its own 4-week average.

New GitHub stars per week for Hermes Agent12 weeks from Aug 9, 2026: 3.8k, 3.5k, 3.9k, 4k, 3k, 2.4k, 2.2k, 2k, 2k, 2.4k, 2.3k, 2.7k new stars, an average of 2.9k a week.Aug 9week of Oct 25
New GitHub stars per week for the last 12 weeks, from GitHub's stargazer history (today's stargazers by the week they starred).
Show the weekly numbers
New GitHub stars per week for Hermes Agent
Week ofNew stars
Aug 9, 20263.8k
Aug 16, 20263.5k
Aug 23, 20263.9k
Aug 30, 20264k
Sep 6, 20263k
Sep 13, 20262.4k
Sep 20, 20262.2k
Sep 27, 20262k
Oct 4, 20262k
Oct 11, 20262.4k
Oct 18, 20262.3k
Oct 25, 20262.7k

On Hacker News

The Board ranks tools by four signals each week: new stars, new stars for the repo's size, added downloads and download growth beyond the market. How the ranking works.

Hermes Agent alternatives

Other AI coding agents on the Board in the week of Oct 25–31, 2026, by momentum:

Coding agents ranked by momentum, the week of Oct 25–31, 2026
RankToolStarsDownloads/wk
#1CindyCindy is an open-source AI agent app for desktop and mobile that combines Claude Code, Codex and Pi across your connected devices.8.8k–
#2OpenMausBotOpenMausBot is an open-source desktop chat app where each contact is a local Claude, Codex, or Grok CLI agent with its own computer.9.6k–
#3OrcaOrca is a desktop agent development environment for running Claude Code, Codex, and other CLI coding agents in parallel, each in its own…111.7k–
#4OpenBotOpenBot is an open-source, self-hosted template for AI coworkers that each get their own browser, files and granted tools, with every…9.1k–
#5Agent ReachAgent Reach is a Python CLI that lets AI agents search and read platforms such as Twitter, Reddit, YouTube, GitHub, Bilibili and…107.2k–
#6Skill RecorderSkill Recorder is a desktop app that records your on-screen work session and turns it into a reusable agent Skill or Automation.5.6k–

All AI coding agents, ranked

Recent Hermes Agent releases

Hermes Agent FAQ

Is Hermes Agent free and open source?

Yes. Hermes Agent's source code is on GitHub (NousResearch/hermes-agent) under the MIT license, so you can use it for free.

How do I install Hermes Agent?

Run curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash. The hermes-agent package needs Python 3.11 or newer, below 3.14. The install section above has the other options from the project's README.

Is Hermes Agent still maintained?

Yes. The NousResearch/hermes-agent repository was last updated on Oct 22, 2026, and the latest release, v0.22.4 on GitHub, came out on Oct 21, 2026.

How popular is Hermes Agent?

Hermes Agent has 259.2k GitHub stars as of Nov 1, 2026. It gained 2.7k new stars in the week of Oct 25–31, 2026, against a 12-week average of 2.9k a week. Its main package had 35.7k downloads that week, −59% against its own 4-week average. That made it #77 among AI coding agents on the Devtoolwatch Board.

Hermes Agent vs OpenCode: how do they compare?

Hermes Agent and OpenCode are both tracked by Devtoolwatch as AI coding agents. Hermes Agent has 259.2k GitHub stars and OpenCode has 223.6k. OpenCode is an open source AI coding agent that helps you write and run code in a terminal, desktop app, or IDE extension. OpenCode review.

Hermes Agent vs Claude Code: how do they compare?

Hermes Agent and Claude Code are both tracked by Devtoolwatch as AI coding agents. Hermes Agent has 259.2k GitHub stars and Claude Code has 154.1k. In the week of Oct 25–31, 2026, Hermes Agent's main package had 35.7k downloads and Claude Code's had 11.6M. Claude Code is Anthropic's agentic coding tool that reads your codebase, edits files, runs commands, and handles git workflows through natural language. Claude Code review.

What are the best alternatives to Hermes Agent?

Other AI coding agents on the Devtoolwatch Board in the week of Oct 25–31, 2026, by momentum: Cindy (#1), OpenMausBot (#2), Orca (#3), OpenBot (#4), Agent Reach (#5). The alternatives table above has their numbers.

Is Hermes Agent local?

Hermes Agent is self-hosted. It can run locally, but also on a VPS, a GPU cluster, SSH hosts, Docker, or serverless backends like Modal and Daytona that hibernate when idle. Hosted agents are managed separately in Nous Portal.

How do I use Hermes Agent with Discord?

Hermes Agent connects to Discord through its messaging gateway. Run hermes gateway setup and hermes gateway start, then send the bot a message. Many slash commands work in both the CLI and messaging platforms.

Does Hermes Agent keep data between sessions?

Hermes Agent stores conversations, memories, and skills so you can return to work later. Stopping a local agent does not delete them. Back up the Hermes data directory when moving to another computer.

Is Hermes Agent safe to use?

Hermes Agent documents security features including command approval, DM pairing, and container isolation. It supports sandboxed backends like Docker and Singularity with container hardening and namespace isolation.

How do I use Hermes Agent for free?

Hermes Agent itself is free under the MIT license. Nous Portal has a $0 tier limited to free models, and a fresh desktop install opens on a guided setup chat running on a free trial model.

Sources

  1. NousResearch/hermes-agent on GitHub
  2. Hermes Agent README
  3. hermes-agent on PyPI
  4. Hermes Agent website
  5. Release v0.22.4 (Oct 21, 2026)
  6. Release v0.22.3 (Oct 21, 2026)
  7. Release v0.22.2 (Oct 15, 2026)
  8. Hacker News: Migrate from OpenClaw
  9. Hacker News: Hermes Desktop
  10. Hacker News: Nous Research edits GitHub issue to remove plagiarism claims about Hermes Agent
  11. Devtoolwatch Board, the week of Oct 25–31, 2026

Hermes Agent README badge

Maintainers can show the tool's weekly Board rank in their README:

Devtoolwatch badge for Hermes Agent

[![Devtoolwatch](https://index.devtoolwatch.com/badge/nousresearch-hermes-agent.svg)](https://index.devtoolwatch.com/tools/nousresearch-hermes-agent/)

How this page is made: Devtoolwatch writes each review from the project's README, releases, package metadata and website, with an AI model, then checks every claim against the source it cites and drops what the sources don't support. The numbers come from Devtoolwatch's daily GitHub snapshots, GitHub's stargazer history, and the npm and PyPI download APIs. Spotted an error? Email [email protected].