fathah/hermes-desktop

Hermes Desktop – A GUI for the Hermes AI Assistant

What it is – Hermes Desktop is a native, cross‑platform (Windows, Linux, macOS) Electron application that wraps the Hermes Agent – a self‑improving, tool‑using LLM assistant created by Nous Research. The desktop client handles the whole lifecycle: installing the agent, configuring LLM providers, managing profiles, and providing a rich chat UI.

Key capabilities

• Guided installation – First‑run wizard runs the official Hermes install script, resolves dependencies (Git, Python 3.11+, uv), and stores the agent under ~/.hermes.
• Local or remote back‑end – Either run the agent locally on 127.0.0.1:8642 or point the app at a remote Hermes API server (URL + API key).
• Multi‑provider LLM support – OpenRouter, Anthropic, OpenAI, Google Gemini, xAI Grok, Nous Portal, Qwen, MiniMax, Hugging Face, Groq, plus any OpenAI‑compatible endpoint (Ollama, LM Studio, vLLM, llama.cpp, etc.).
• Streaming chat UI – Server‑sent‑events (SSE) feed live token counts, markdown, syntax‑highlighted code, and tool progress bars.
• Slash‑command toolbox – 22 built‑in commands (/new, /clear, /web, /image, /code, /shell, /memory, /persona, /schedule, …) let you control the session, invoke tools, or query usage.
• Session & profile management – SQLite‑backed history with full‑text search, date grouping, and isolated configuration folders for multiple “agents”.
• Tool ecosystem – 14 ready‑to‑use toolsets (web browsing, terminal, file ops, vision, image generation, TTS, skill execution, memory providers, delegation, task planning, etc.).
• Memory system – View/edit entries, track capacity, and plug in providers such as Honcho, Hindsight, Mem0, RetainDB, Supermemory, ByteRover.
• Persona editor – Edit the agent’s SOUL.md personality directly from the UI.
• Scheduling – Cron‑style job builder with 15 delivery targets (e.g., messaging gateways).
• Messaging gateways – Integrated bots for Telegram, Discord, Slack, WhatsApp, Signal, Matrix, Mattermost, Email, SMS, iMessage (BlueBubbles), DingTalk, Feishu/Lark, WeCom, WeChat, Webhooks, Home Assistant.
• Hermes Office (Claw3d) – A 3‑D visual interface for managing adapters and dev server.
• Backup / import / diagnostics – One‑click data export, restore, and log viewer.
• Auto‑updates – Electron‑updater checks GitHub releases and applies patches.
• Internationalisation – English UI ready for community translations.
• Test suite – Vitest‑based unit tests for SSE parsing, IPC, installer utilities, etc.

How it works

1. On first launch you choose local or remote mode.
2. Local: the app runs the Hermes install script, places the agent in ~/.hermes, and starts the HTTP server.
3. Remote: you supply the remote URL and API key; the app skips installation.
4. You then select an LLM provider or a local OpenAI‑compatible endpoint; the credentials are stored in ~/.hermes/.env (or via a configurable “command” secret provider).
5. Chat requests are sent to the chosen backend using SSE; the client parses the stream and updates the UI in real time.

Who might use it – Developers or power users who want a desktop‑friendly interface for a self‑hosting LLM assistant, especially those who need multi‑provider flexibility, tool integration, or messaging‑gateway bots without dealing with CLI configuration.

Installation – Downloadable installers for Windows (unsigned) and Fedora RPMs are provided. The README includes notes for WSL, unsigned installers, and sudo‑less install tricks.

Community – The project is open‑source (MIT), accepts contributions, and has a Discord community. Sponsorship options and a Ko‑fi link are listed.

---

All details above are taken directly from the repository’s README.