Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/techjarves/Hermes-USB-Portable/llms.txt

Use this file to discover all available pages before exploring further.

This page walks you through launching Hermes USB Portable for the first time — from copying the folder to your drive all the way to your first chat session. The launcher handles all dependency downloads automatically, so you only need to follow a few steps per platform.

​
Prerequisites

No programming tools need to be installed on the host computer. Everything is downloaded and managed automatically on first run.
  • A USB flash drive, external SSD, or any local folder with approximately 2 GB of free space (4 GB recommended for comfortable headroom).
  • An internet connection for the first-run setup, which downloads roughly 600 MB of runtimes and source code. Subsequent launches work entirely offline.

​
Launch the Portable Folder

The simplest method is to double-click launch.bat in the portable folder. Windows will open a terminal window automatically.Alternatively, open PowerShell in the portable folder and run:
.\launch.bat
On first run, the launcher opens a PowerShell session to download dependencies. If Windows SmartScreen appears, click More info → Run anyway. The scripts are fully open-source and human-readable in the scripts/ directory.

​
First-Run Setup

On the very first launch the setup script runs automatically before the menu appears. It downloads and installs everything Hermes needs — no manual steps required.
1

Portable Python 3.11

A self-contained CPython build from python-build-standalone is downloaded to .cache/runtimes/<platform>-<arch>/python/. This is the only Python interpreter Hermes will ever use.
2

Node.js 22 LTS

A portable Node.js binary is downloaded to .cache/runtimes/<platform>-<arch>/node/ to support Hermes web and JavaScript tools.
3

uv, ripgrep, and tooling

The uv package manager and ripgrep are downloaded to the runtime directory and used for all subsequent package operations.
4

Hermes Agent source code

The Hermes Agent source tree is cloned from NousResearch/hermes-agent into src/hermes-agent/.
5

Python virtual environment and dependencies

An isolated virtual environment is created using uv, and all Hermes Python dependencies (including Playwright/Chromium for web tools) are installed locally. This step may take 3–10 minutes depending on your connection speed.
6

Ready flag written

A ready.flag file is created in the runtime directory. All future launches skip setup and go directly to the menu in seconds.
The total first-run download is approximately 600 MB. The full setup (including archives, venv, and Playwright browsers) measures around 1.5 GB on Windows x64. Make sure your drive has sufficient free space before starting.

​
Using the Terminal Dashboard

After setup completes (or on any subsequent launch), the interactive terminal dashboard appears: 
----------------------------------------------------------------
                    HERMES PORTABLE LAUNCHER
                         AI Agent for Everyone
----------------------------------------------------------------

 Setup    [OK] Configured
 Provider openrouter
 Model    nous/hermes-3-70b
 Gateway  [ ] Stopped
 Version  v1.2.0

----------------------------------------------------------------

  [1]  Start Hermes Chat
  [2]  Setup / Reconfigure Hermes
  [3]  Start Gateway
  [4]  Advanced Options  -->
  [5]  Exit

----------------------------------------------------------------
OptionAction
[1] Start Hermes ChatOpens an interactive AI chat session in the terminal.
[2] Setup / ReconfigureRuns the Hermes setup wizard to configure your LLM provider and API keys. Start here on a fresh install.
[3] Start/Stop GatewayStarts or stops the background messaging gateway (Telegram, etc.). The label toggles to Stop Gateway when a gateway process is live.
[4] Advanced OptionsSub-menu with Run Doctor, View Logs, Edit Config, Restart Gateway, and Update Hermes.
On first use, select [2] Setup / Reconfigure Hermes to enter your API key and choose your LLM provider before starting a chat session.

​
Next Steps

Configure API Keys

Add your OpenRouter, OpenAI, Anthropic, or other provider credentials to data/.env.

Launcher Menu Guide

Explore every option in the terminal dashboard, including advanced options and gateway management.

Local Ollama Setup

Run Hermes entirely offline using a locally hosted Ollama model — no API key required.

Build docs developers (and LLMs) love

Get started for freeTalk to us