Guide · Open source
How to use hermes-desktop: install, set up and chat
Updated · written with AI from the project's own documentation and checked against our catalogue data
Hermes Desktop is a community-maintained desktop app that installs, configures and chats with Hermes Agent, an AI assistant with tool use and messaging integrations. It is aimed at people who want Hermes without working in the command line. This guide walks through installing the app, the first-run setup, and the main areas you will use day to day.
Quick start
- 1Download the app. Get the installer for your operating system from hermesone.org or GitHub Releases.
- 2Install it. Run the installer. On Windows, click "More info" and then "Run anyway" if SmartScreen warns you.
- 3Let first-run setup finish. The app installs Hermes Agent using the official install script and stores data in ~/.hermes.
- 4Connect a provider. Choose a provider such as OpenRouter, Anthropic or a local Ollama server, and add your credentials or endpoint.
- 5Start chatting. Open a new session and send a message, or use slash commands like /new and /model.
- 6Add profiles or automations. Create a separate profile for other setups, or build a scheduled task or messaging gateway.
Download and install Hermes Desktop
Download the desktop app for your operating system from hermesone.org or GitHub Releases. The README says that npm install hermes-desktop is not a supported desktop installation method.
On Fedora, install the downloaded RPM with the command below, replacing <version> with the version of the file you downloaded.
- Windows: the installer is not code-signed, so SmartScreen will warn on first launch. Click "More info", then "Run anyway".
- WSL: if the installer stalls at "Switching to root user to install dependencies...", the README explains a temporary passwordless sudo workaround. Remove it when the install finishes.
- Fedora: the .rpm is not GPG-signed. If your system enforces signature checking, the README says to add --nogpgcheck.
sudo dnf install ./hermes-desktop-<version>.rpmRun the first-run setup
On first launch, the app installs Hermes Agent for you. The README says it uses the official Hermes install script and stores Hermes data in ~/.hermes. Features listed for the guided install include progress tracking and dependency resolution.
You can run Hermes on your own machine at 127.0.0.1:8642, or point the desktop app at a remote Hermes API server by entering its URL and API key.
Connect a model provider
Provider setup is handled in the app. The README lists OpenRouter, Anthropic, OpenAI, Google (Gemini), xAI (Grok), Nous Portal, Qwen, MiniMax, Hugging Face and Groq, plus local OpenAI-compatible endpoints such as LM Studio, Atomic Chat, Ollama, vLLM and llama.cpp.
Saved models let you create, edit and delete model configurations across providers, so you can switch between them.
Chat, use slash commands and manage sessions
The chat window streams responses and shows tool progress, markdown rendering and syntax highlighting. The chat footer shows live prompt and completion token counts and cost, and the /usage slash command reports usage too.
Sessions are searchable with full-text search (SQLite FTS5) and grouped by date, so you can resume an earlier conversation. The app lists 22 slash commands, including /new, /clear, /model, /memory and /skills.
Use profiles, memory, skills and the persona editor
Profiles let you create, delete and switch between separate Hermes environments, each with its own config. This is useful if you want to keep several setups apart.
The memory screen lets you view and edit memory entries and your user profile memory, and shows capacity. The persona editor edits and resets the SOUL.md personality file for your agent.
- Toolsets: 14 groups, including web, browser, terminal, file, code execution and vision.
- Discover: browse and add skills from the Discover screen.
Set up scheduled tasks and messaging gateways
The scheduler is a cron job builder with options for minutes, hourly, daily, weekly and custom cron, plus 15 delivery targets.
The gateway screen supports 16 messaging platforms, including Telegram, Discord, Slack, WhatsApp, Signal, Matrix and Email. Each one is set up through the graphical interface rather than by editing config files by hand.
What people use it for
- Chatting with Hermes Agent without using the terminal.
- Switching between cloud providers and a local Ollama or LM Studio server from one window.
- Keeping separate Hermes profiles for work and personal use.
- Connecting Telegram or Discord to your agent through the gateway screen.
- Running scheduled tasks with the cron job builder.
Frequently asked questions
Is Hermes Desktop free?
The project is released under the MIT licence, as stated in its README. Check the licence file in the repository for the full terms. Any model provider you connect may charge its own fees.
Do I need to install anything before Hermes Desktop?
The source lists no prerequisites for the packaged app. You download the installer for your operating system and run it. Developers who want to run from source should follow the Contributing Guide.
Why does Windows show a SmartScreen warning?
The Windows installer is not code-signed, so SmartScreen warns on first launch. Click "More info" and then "Run anyway" to continue. Only do this if you trust the download source.
Can I update Hermes Desktop on Fedora automatically?
No. The README says auto-update is not supported for .rpm builds. Reinstall the new .rpm to update.
Can I use Hermes Desktop with Ollama or LM Studio?
Yes. The README lists local OpenAI-compatible endpoints including Ollama and LM Studio as supported provider options.
Is Hermes Desktop stable enough for daily use?
The README says the project is in active development, features may change, and some things might break. Report problems through its GitHub issues page.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Hermes Desktop suits people who want Hermes Agent with a graphical interface, multiple profiles, and provider or messaging setup without the command line. Developers wanting a stable product, or Fedora users who need automatic updates, should look elsewhere or use the Hermes Agent CLI directly. Comparable options named in the source include Open WebUI, LM Studio and Jan.
Install options & downloads