HuggingMes
somratpro/HuggingMes
Run Hermes Agent as an always-on assistant on Hugging Face Spaces with dashboard, backup and Telegram
HuggingMes is a self-hosted deployment of Hermes Agent for Hugging Face Spaces that aims to keep a personal assistant running around the clock. It adds a management dashboard, backup of chats and config to a private HF Dataset, and Cloudflare Worker fixes for blocked outbound traffic.
What HuggingMes does
The project runs Nous Research's Hermes Agent inside a Docker Space on Hugging Face. A gateway token protects the dashboard and API, and the dashboard at / shows uptime, sync health, model, provider and agent status. Chats, config and session data sync to a private HF Dataset. A Cloudflare Worker can proxy Telegram and other outbound traffic that the Space blocks, or keep the Space awake with a cron trigger. The gateway, dashboard, health server and JupyterLab are monitored and restarted if they exit.
A JupyterLab terminal is enabled at /terminal/ once GATEWAY_TOKEN is set. Shell wrappers record apt, pip, uv, npm and hermes installs and replay them on every boot so they survive restarts, and startup scripts can run through HUGGINGMES_RUN. Comma-separated key pools such as ANTHROPIC_API_KEYS rotate keys. It works with Hermes' native providers, direct API-key providers, OAuth providers and custom OpenAI-compatible endpoints.
Key features
- Hermes Agent running in a Hugging Face Docker Space
- Dashboard with uptime, sync health, model and provider status
- Backup of chats, config and sessions to a private HF Dataset
- Cloudflare Worker proxy for Telegram and a keep-alive cron trigger
- JupyterLab terminal at /terminal/ and replay of package installs on boot
- API key pool rotation
When to use it
- Host a Telegram-connected Hermes assistant without your own server
- Keep chats and config backed up while the Space restarts
- Reach a terminal on the hosted agent through JupyterLab
Who it is for: Hermes Agent users who want to host an always-on assistant on Hugging Face Spaces instead of a VPS.
How it fits with Hermes Agent
Built around Hermes Agent: it packages and runs Hermes on Hugging Face Spaces and can use providers configured through hermes model or config.yaml.
Requirements: A Hugging Face account and Space with the secrets listed in the README: LLM_API_KEY, LLM_MODEL, GATEWAY_TOKEN, TELEGRAM_BOT_TOKEN, TELEGRAM_ALLOWED_USERS, HF_TOKEN and CLOUDFLARE_WORKERS_TOKEN.
Note: The README warns that using this project may lead to suspension of your Hugging Face account.
FAQ
What is HuggingMes?
HuggingMes is a way to run Hermes Agent on Hugging Face Spaces as a 24/7 personal assistant. It adds a dashboard, persistent backup to an HF Dataset and connectivity fixes.
Does HuggingMes work with Hermes Agent?
Yes, it runs Nous Research's Hermes Agent inside the Space. It wires several providers directly and can use others configured through hermes model or config.yaml.
What do I need to run HuggingMes?
You need a Hugging Face account and Space plus secrets such as an LLM API key, a model name and a gateway token. Telegram, HF_TOKEN and Cloudflare Workers tokens are also listed in the README.
Similar deployment for Hermes Agent
All deploymentStreamlit script builder for Fedora Workstation that can install Hermes Agent among other tools
clawfleet ClawFleetDocker-based fleet manager that runs OpenClaw and Hermes agents from one browser dashboard
nickvasilescu Nick's StackReady-to-run Hermes agent template with Telegram onboarding and bring-your-own keys
ultraworkers Unofficial Hermes Agent Helm ChartCommunity Helm chart that packages Hermes Agent for Kubernetes with state-safety guardrails
sunnysktsang Hermes SuiteOne container running Hermes' gateway, dashboard and web UI with auto runtime detection
iranee Hermes Agent for fnOSNative fnOS NAS package that deploys Hermes Agent with a web control panel
Related guides: How to run Hermes Agent securely · How to install Hermes Agent · Connect Hermes agents on several machines with Hermes Desktop