A lightweight self-hosted AI agent written in Go that runs on your server and works through Telegram, using Anthropic, OpenAI or OpenRouter models.
imkofty/Hermuse
Scripts and guides to run Hermes Agent, the 9Router model proxy and a Telegram bot gateway on your own Linux server or VM.
- Stars
- 157
- Forks
- 39
- Open issues
- 0
- Last commit
- 16 days ago
What is it?
Hermuse is a set of shell scripts and documentation for running a personal AI agent stack on your own machine or a VPS. It combines the Hermes Agent CLI, the 9Router OpenAI-compatible model proxy, and a Telegram gateway, so you can chat with your own AI bot around the clock and open a model dashboard from any browser.
The repository covers installation on Ubuntu, a Cloudflare tunnel that exposes the local dashboard through Cloudflare Pages and D1 using a polling client, and supervisors that restart 9Router, the tunnel client and the gateway when they fail. A watchdog script runs from cron, and a separate gateway watchdog detects stalled Telegram polling rather than only checking that the process exists.
The project is aimed at Linux VMs and ships a doctor script, an allowlist-based access model for the Telegram bot, and guides for credential rotation and tunnel database migrations.
Key features
- 01Installs Hermes Agent, 9Router and Node.js on Ubuntu with a single script.
- 02Connects a Telegram bot to the Hermes gateway with a numeric user ID allowlist.
- 03Exposes the local 9Router dashboard publicly through a Cloudflare Pages and D1 polling tunnel.
- 04Auto-restarts 9Router, the tunnel client and the gateway via supervisor scripts and cron watchdogs.
- 05Detects stalled Telegram polling using an event-loop heartbeat and adapter activity.
- 06Provides a doctor script and a migration checker for the tunnel database.
Use cases
- Running a private Telegram AI assistant on a small Linux VPS.
- Hosting a self-managed model dashboard reachable from a browser without opening inbound ports.
- Operating an always-on bot that needs automatic recovery after crashes or reboots.
Quick start
- 1
Install everything (dependencies, Hermes, 9Router)
bash scripts/install.sh - 2
Log in to the model provider via OAuth
hermes setup --portal - 3
Connect your Telegram bot and set the numeric-ID allowlist
hermes gateway setup - 4
Install the Cloudflare tunnel (replace the placeholders with your project and D1 names)
bash scripts/setup-tunnel.sh <pages-project-name> <d1-name>
Things to know
- Designed for a Linux server or VM; the README gives no macOS or Windows setup.
- Requires passwordless sudo or root for the install script.
- Needs a Cloudflare account and a Telegram bot token to use the full stack.
In collections
Topics
FAQ
- What is Hermuse?
- Scripts and guides to run Hermes Agent, the 9Router model proxy and a Telegram bot gateway on your own Linux server or VM. Hermuse is a set of shell scripts and documentation for running a personal AI agent stack on your own machine or a VPS. It combines the Hermes Agent CLI, the 9Router OpenAI-compatible model proxy, and a Telegram gateway, so you can chat with your own AI bot around the clock and open a model dashboard from any browser.
- Is Hermuse free to use? What license does it have?
- Hermuse is open source under the MIT license. Check the license terms before using it in commercial products.
- Is Hermuse still maintained?
- Actively maintained. The last commit to Hermuse was on Oct 23, 2026.
- What can I use Hermuse for?
- Running a private Telegram AI assistant on a small Linux VPS. Hosting a self-managed model dashboard reachable from a browser without opening inbound ports. Operating an always-on bot that needs automatic recovery after crashes or reboots.
- How do I install Hermuse?
- Install everything (dependencies, Hermes, 9Router): bash scripts/install.sh Log in to the model provider via OAuth: hermes setup --portal Connect your Telegram bot and set the numeric-ID allowlist: hermes gateway setup Install the Cloudflare tunnel (replace the placeholders with your project and D1 names): bash scripts/setup-tunnel.sh <pages-project-name> <d1-name>
- Where can I get Hermuse?
- The source code and installation instructions are on GitHub: https://github.com/imkofty/Hermuse
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A self-hosted personal AI agent that chats with you and, after approval, carries out local tasks like file work, commands and scheduled jobs.
Self-hosted AI personal assistant that chats via Telegram, runs scheduled tasks, and connects to MCP servers and host tools.
A self-hosted platform of persistent AI agents with long-term memory that collaborate and answer across chat channels, run in a single container.
A personal AI agent that runs as a control plane in your own Cloudflare account and acts across your laptop, server, phone and browser.
A Rust-based personal AI agent runtime you run on your own machine, connecting LLM providers to chat channels, tools and hardware.