Hermes One is a community-maintained desktop application for Hermes Agent, the self-improving assistant from Nous Research. It replaces hands-on management of the Hermes CLI with a guided first-run setup and a graphical workspace for chat, sessions, profiles, memory, skills, tools, scheduled tasks, and messaging gateways. Hermes itself is installed with the official install script into ~/.hermes.
On first launch the app asks whether to run Hermes locally or connect to a remote Hermes API server. Local mode checks for an existing install, resolves dependencies such as Git, uv, and Python 3.11+, and runs the installer; remote mode takes a URL and API key and skips installation. It then prompts for a model provider or local endpoint, saves the configuration through Hermes config files, and opens the main workspace. Chat requests go to 127.0.0.1:8642 (or the remote URL) over SSE, and the app renders tool progress, Markdown, and token usage as the stream arrives.
The project is in active development; features may change between releases. Installers are provided for Windows, macOS, and Linux, including an RPM build for Fedora.
Features
- Guided install: first-run flow with progress tracking and dependency resolution, or connection to a remote Hermes API server
- Providers: OpenRouter, Anthropic, OpenAI, Google Gemini, xAI, Nous Portal, Qwen, MiniMax, Hugging Face, Groq, and local OpenAI-compatible endpoints such as LM Studio, Ollama, vLLM, and llama.cpp
- Streaming chat: SSE streaming with tool progress indicators, Markdown rendering, syntax highlighting, and live token and cost counts
- Slash commands: 22 commands including /new, /web, /browse, /code, /shell, /model, /memory, /persona, /compact, /undo, and /retry
- Sessions and profiles: SQLite FTS5 full-text search across conversation history, and isolated Hermes environments you can create and switch between
- Toolsets: 14 groups covering web, browser, terminal, files, code execution, vision, image generation, TTS, skills, memory, delegation, MoA, and task planning
- Memory: view and edit memory entries and user profile memory, with discoverable providers such as Honcho, Hindsight, Mem0, and Supermemory
- Scheduling: a cron job builder with minute, hourly, daily, weekly, and custom schedules and 15 delivery targets
- Messaging gateways: 16 channels including Telegram, Discord, Slack, WhatsApp, Signal, Matrix, email, SMS, iMessage, Feishu, WeCom, webhooks, and Home Assistant
- Persona and models: a SOUL.md persona editor and saved model configurations per provider
- Maintenance: backup and restore, debug dumps, a log viewer, and updates through electron-updater
Integrates with
Firecrawl
Web data API that searches, scrapes, and crawls sites into clean markdown or structured JSON for agents
Hermes
Self-improving AI agent with a learning loop that creates and refines skills, recalls past sessions, and runs across chat platforms
Hindsight
Agent memory system that stores and retrieves long-term memories so agents learn across sessions, not just recall history
llama.cpp
LLM inference in C/C++ across CPU and GPU backends, using the GGUF format with quantization, a REST server, and a WebUI