How to Set Up Hermes Agent: Complete Beginner's Guide (2026)

TL;DR: Hermes Agent is an open-source AI agent by Nous Research that learns and improves over time. Install it in under 2 minutes with a one-line command, connect it to Telegram or Discord, and it starts building skills from your conversations. This guide walks you through everything from installation to advanced features like autonomous skill creation and multi-platform messaging.

---

What Is Hermes Agent?

Hermes Agent isn't another AI chatbot wrapper. It's a self-improving agent built by Nous Research — the lab behind the Hermes, Nomos, and Psyche model families. Released in February 2026, it hit 70,000+ GitHub stars in its first three months.

What makes it different from other agents (OpenClaw, Claude Code, Copilot) is the learning loop. Hermes doesn't just execute tasks. It:

• Creates reusable skills from complex tasks it completes
• Improves those skills during subsequent use
• Nudges itself to persist knowledge across sessions
• Searches its own past conversations with full-text search
• Builds a deepening model of who you are over time

Think of it as the difference between an assistant that remembers your name and a partner that actually gets better at helping you the more you work together.

---

Before You Start

What You'll Need

• A computer running Linux, macOS, or WSL2 on Windows (Android via Termux also works)
• Git — the only prerequisite you need to install yourself
• An API key or subscription for an AI model provider (OpenAI, Anthropic, OpenRouter, or one of 20+ supported options)
• About 10 minutes to get through the full setup

Supported Platforms

Platform	Support
Linux (any distro)	Full
macOS (Intel & Apple Silicon)	Full
Windows	Via WSL2
Android	Via Termux
NixOS	Full (Nix flake available)

---

Step 1: Install Hermes Agent

The installer is a single curl command:

curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash

What the Installer Does

It handles everything automatically:

1. Installs Python 3.11 (via uv — fast Python package manager)
2. Installs Node.js v22 (for browser automation and WhatsApp bridge)
3. Installs ripgrep (for fast file search in memory)
4. Installs ffmpeg (for audio format conversion and TTS)
5. Clones the Hermes Agent repo to ~/.hermes/hermes-agent/
6. Creates a Python virtual environment and installs dependencies
7. Sets up the hermes command globally

You don't need to install Python, Node.js, or any of those tools manually. The only prerequisite is Git.

After Installation

Reload your shell:

source ~/.bashrc # or: source ~/.zshrc

Then verify it's installed:

hermes --version

---

Step 2: Choose a Model Provider

This is the most important configuration step. Run:

hermes model

This launches an interactive menu. The best options for most users:

Provider	Best For	Setup
Nous Portal	Zero-config, subscription-based	OAuth login (easiest option)
OpenAI Codex	Codex models via ChatGPT OAuth	Device code auth
Anthropic	Claude models	OAuth (Max plan) or API key
OpenRouter	200+ models, pay-per-token	Enter your API key
Custom Endpoint	Local models (Ollama, vLLM)	Set base URL + API key

Important: Hermes requires a model with at least 64K tokens of context. Most hosted models (Claude, GPT, Gemini, Qwen, DeepSeek) meet this easily. If running a local model, make sure your context size is set to 64K or higher.

You can switch providers at any time with hermes model — no lock-in, no code changes.

---

Step 3: Run Your First Chat

Start the CLI:

hermes # Classic CLI
hermes --tui # Modern TUI (recommended)

The TUI (terminal user interface) features:

• Multiline editing
• Slash-command autocomplete
• Conversation history
• Interrupt-and-redirect mid-response
• Streaming tool output

Try a prompt like:

Summarize this repo in 5 bullets and tell me what the main entrypoint is.

Or:

Check my current directory and tell me what's the main project file.

Quick Verification

Run this command to make sure everything is wired up correctly:

hermes doctor

This checks your installation, config, provider connectivity, and tool availability in one go.

---

Step 4: Set Up Messaging (Gateway)

Once the CLI works, you'll want to talk to Hermes from your phone. The gateway connects Hermes to your messaging platforms:

hermes gateway setup

This walks you through connecting:

• Telegram — Create a bot via BotFather, paste the token
• Discord — Create a bot application, configure intents
• Slack — Install the Hermes Slack app to your workspace
• WhatsApp — Via the WhatsApp Business API bridge
• Signal — Signal CLI integration

After setup, start the gateway:

hermes gateway

Your agent is now always-on. Message it from any connected platform.

---

Step 5: Configure Tools

Hermes ships with 40+ built-in tools. See what's available:

hermes tools

Key tool categories:

• Web Search & Browse — Search, extract, and browse web pages
• File System — Read, write, edit files in your workspace
• Code Execution — Run Python, shell commands, and scripts
• Image Generation — Generate images via supported providers
• Browser Automation — Headless browser for web tasks
• TTS / Voice — Text-to-speech and voice memo transcription
• MCP Support — Connect to any MCP server for extended capabilities

Enable only what you need. You can enable/disable individual tools at any time.

---

Step 6: The Learning Loop — Skills & Memory

This is where Hermes differentiates itself from every other agent.

Autonomous Skill Creation

After you complete a complex workflow with Hermes, ask it to save what it learned:

Save that as a skill for next time.

Hermes will:

1. Analyze what it just did
2. Create a reusable skill script
3. Store it in ~/.hermes/skills/
4. Load it automatically in future conversations

How Skills Work

Skills are portable, shareable Python scripts. They follow the agentskills.io open standard. You can:

• Browse your skills: hermes skills list
• Share them with the community via the Skills Hub
• Let Hermes improve them during use — skills get better over time

Memory System

Hermes uses FTS5 full-text search across all past sessions with LLM summarization for cross-session recall. It also uses Honcho dialectic user modeling — a technique that builds a deepening model of your preferences, work style, and context.

The agent periodically nudges itself to persist important knowledge. You don't have to tell it to remember — it does that autonomously.

---

Advanced Setup

Serverless Deployment

Hermes supports six terminal backends for how and where it runs:

hermes config set terminal.backend local # your machine
hermes config set terminal.backend docker # Docker container
hermes config set terminal.backend ssh # remote server
hermes config set terminal.backend daytona # serverless (hibernates when idle)
hermes config set terminal.backend modal # serverless GPU
hermes config set terminal.backend singularity # HPC clusters

Daytona and Modal are game-changers. Your agent's environment hibernates when idle and wakes on demand — costing nearly nothing between sessions. Run Hermes on a $5 VPS and it works like a cloud service.

Scheduled Automations (Cron)

Hermes has a built-in cron scheduler:

hermes cron add --schedule "0 9 * * 1" --task "Summarize last week's notes"

Tasks deliver to any connected messaging platform. Use it for:

• Daily research briefings
• Weekly code review summaries
• Nightly backups
• Periodic audit reports

Multi-Provider Fallback

Configure fallback providers so if one model is down, another takes over:

hermes config set model anthropic/claude-sonnet-4
hermes config set fallback openai/gpt-5.1-codex

---

Migrating from OpenClaw

If you're coming from OpenClaw, Hermes ships with a migration tool:

hermes claw migrate

This imports your OpenClaw configuration, tools, and workflows. Most settings transfer automatically — you'll be up and running in minutes.

---

Common Troubleshooting

Problem	Solution
hermes: command not found	Reload your shell (source ~/.bashrc) or check your PATH
API key not set	Run hermes model to configure your provider
Model context too small	Set at least 64K context (models with smaller windows won't work)
Gateway won't start	Check hermes doctor for configuration issues
Skills not loading	Run hermes skills list to verify they're installed
After update, something breaks	Run hermes config check then hermes config migrate

---

Next Steps

Now that you have Hermes running:

1. Explore the TUI — hermes --tui with its modal overlays and mouse support
2. Connect Telegram — So you can message your agent from anywhere
3. Create your first skill — Complete a multi-step task and ask Hermes to save it
4. Set up cron — Automate your first recurring task
5. Browse the Skills Hub — Install community-contributed skills

Hermes Agent is one of those rare tools that compounds in value. The more you use it, the better it gets — and the more it learns about how you work, the more autonomously it can handle things for you.

---

Quick Reference

# Installation
curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash
Commands
hermes # Start chatting
hermes —tui # Modern TUI interface
hermes model # Choose your AI provider
hermes tools # Configure tools
hermes gateway setup # Set up messaging
hermes gateway # Start messaging gateway
hermes config set # Set individual config values
hermes setup # Full setup wizard
hermes skills list # Browse your skills
hermes cron add # Schedule automation
hermes doctor # Diagnose issues
hermes claw migrate # Migrate from OpenClaw

---

Tags: Guides, AI, Open Source, Tutorials

Published May 2026 on ToolBrain

← Back to all posts