Nous Research Launches Official Hermes Desktop

Hermes Agent

In recent years, the AI Agent field has developed rapidly, evolving from simple chatbots into agents capable of truly executing tasks, maintaining persistent memory, and growing autonomously.

Hermes Agent is an open-source autonomous AI Agent framework by Nous Research. Through natural language interaction, this autonomous AI Agent can execute terminal commands, control browsers, read and write files, search the internet, generate images, and maintain memory across sessions.

- GitHub Repository: https://github.com/nousresearch/hermes-agent

- Hermes Agent Official Site: https://hermes-agent.nousresearch.com/

The official definition is very direct:

The agent that grows with you. An Agent that grows continuously with use. It is an autonomous agent; the longer it runs, the stronger its capabilities become.

The core philosophy of Hermes Agent is: To make AI a long-term online digital employee, rather than a one-time chatbot.

Hermes Agent is one of the few AI Agents in the industry with a native built-in learning loop. It can accumulate skills from execution experience, autonomously optimize capabilities, persist knowledge, retrieve historical conversations, and continuously refine the user cognitive model across sessions.

Difference from ordinary AI chat tools:

Core Features:

- Connect: Telegram, Discord, Slack, WhatsApp, Signal, Email, CLI — one Agent, unified memory, all platforms. A conversation started on Telegram can be seamlessly continued in the terminal.

- Remember: Hermes learns your projects, automatically generates skills, and never forgets problems it has solved. After each session, important information is written to persistent memory.

- Schedule: Set scheduled tasks using natural language — daily reports, backups, routine reviews, morning briefings — running unattended in the background.

- Delegate: Generate isolated sub-agents with independent conversation context, independent terminals, and Python RPC scripts, achieving parallel pipelines with zero context cost.

- Search & Multimodal: Web search, browser automation, visual understanding, image generation, text-to-speech, multi-model inference, all built-in.

Hermes Agent supports free switching between any large models, including Nous Portal, OpenRouter (200+ models), OpenAI, GLM, Kimi, MiniMax, etc. Execute `hermes model` to switch, no code changes required, no vendor lock-in.

Of course, buying the Coding Plan is still the most cost-effective, after all, it's monthly:

Overall Architecture

Overall Flow:

Architecture Diagram:

Quick Installation

The following installation commands apply to Linux, macOS, and WSL2 systems:

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

Reload Shell after installation:

source ~/.bashrc # Use bash # Or source ~/.zshrc # Use zsh

Windows Native Installation (PowerShell):

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

Windows native support is still experimental; WSL2 is recommended.

macOS / Windows Desktop Version (Desktop App): Go to hermes-agent.nousresearch.com to download the desktop installation package for the corresponding platform. CLI and desktop application will be installed simultaneously.

Double-click to install after downloading: Verify Installation:

hermes --version

Outputting the version number indicates successful installation.

The installer automatically handles all dependencies:

- uv — Fast Python package manager

- Python 3.11 — Installed via uv, no sudo required

- Node.js v22 — For browser automation and WhatsApp bridging

- ripgrep — Fast file search

- ffmpeg — TTS audio format conversion

Windows System Note: Native Windows environment is not supported. Please install WSL2 first, then execute the above commands in the WSL2 terminal.

If you have installed OpenClaw, you can directly import the configuration:

Here choose not to import, enter n, enter model settings, select the first quick setting:

Next we select the second to last item More providers:

Here you can select more domestic or international large models:

Next we use ByteDance Volcano Engine Coding Plan, so we choose Customer endpoint, which is custom API address and API key.

Base URL setting ByteDance Volcano Engine compatible OpenAI interface protocol tool: https://ark.cn-beijing.volces.com/api/coding/v3

Set API key and supported large models, such as ark-code-latest.

After installation, execute the following command:

source ~/.bashrc # Reload shell config (if using zsh, execute: source ~/.zshrc) hermes # Start agent conversation!

Quick Start

hermes # Interactive command line interface — Start conversation hermes model # Select large language model provider and corresponding model hermes tools # Configure enabled tool set hermes config set # Set single configuration item hermes gateway # Start message gateway (supports Telegram, Discord, etc.) hermes setup # Run full configuration wizard (one-stop completion of all configurations) hermes claw migrate # Migrate data from OpenClaw (for original OpenClaw users) hermes update # Update to latest version hermes doctor # Diagnose runtime environment and configuration issues

Manual Installation

If you prefer complete control over the installation process, please follow the steps below.

Step 1: Clone Repository

Clone with --recurse-submodules to get required submodules:

git clone --recurse-submodules https://github.com/NousResearch/hermes-agent.git cd hermes-agent

If already cloned but without submodules:

git submodule update --init --recursive

Step 2: Install uv and Create Virtual Environment

# Install uv (if not installed) curl -LsSf https://astral.sh/uv/install.sh | sh # Create Python 3.11 virtual environment (uv will download when needed, no sudo) uv venv venv --python 3.11

hermes. The entry point has a hardcoded shebang pointing to Python in the virtual environment, so it can be used after global linking.

Step 3: Install Python Dependencies

# Tell uv which virtual environment to install to export VIRTUAL_ENV="$(pwd)/venv" # Install all extensions uv pip install -e ".[all]"

If you only want the core agent (without Telegram/Discord/cron support):

uv pip install -e "."

Optional Extension Explanation:

We can combine usage: uv pip install -e ".[messaging,cron]"

Step 4: Create Configuration Directory

# Create directory structure

mkdir -p ~/.hermes/{cron,sessions,logs,memories,skills,pairing,hooks,image_cache,audio_cache,whatsapp/session}

# Copy example configuration file

cp cli-config.yaml.example ~/.hermes/config.yaml

# Create empty .env file for storing API keys

touch ~/.hermes/.env

Step 5: Add API Keys

Open ~/.hermes/.env and add at least one LLM provider key:

# Required — At least one LLM provider: OPENROUTER_API_KEY=sk-or-v1-your-key-here # Optional — Enable other tools: FIRECRAWL_API_KEY=fc-your-key # Web search and scraping FAL_KEY=your-fal-key # Image generation (FLUX)

Or set via CLI:

hermes config set OPENROUTER_API_KEY sk-or-v1-your-key-here

Step 6: Add hermes to PATH

mkdir -p ~/.local/bin ln -sf "$(pwd)/venv/bin/hermes" ~/.local/bin/hermes

If ~/.local/bin is not in PATH, add to shell config:

# Bash echo 'export PATH="$HOME/.local/bin:$PATH"' >> ~/.bashrc && source ~/.bashrc # Zsh echo 'export PATH="$HOME/.local/bin:$PATH"' >> ~/.zshrc && source ~/.zshrc

Step 7: Configure Your Provider.

This page provides an editorial summary based on publicly available information. It is not a republished article. Use the source link below for the original report.