GitHub radar
Hermes Agent: A Self-improving AI Agent developed by Nous Research
Hermes Agent by Nous Research is a self-improving AI agent that works 24/7 on low cost cloud services and takes tasks from Telegram, Discord, or Slack while learning from each task it does.
Hermes Agent is ‘The self-improving AI agent built by Nous Research.’ It works 24/7 on low cost cloud services ($5/month VPS, essentially free on serverless platforms like Modal and Daytona [hibernation mode]) and takes tasks from Telegram, Discord, Slack or terminal. Unique features include learning from each task performed and developing reusable skills to speed up future similar tasks, as well as building a persistent user identity. Compatible with any LLM backend (Claude, OpenAI, Nous Portal, self-hosted). 7 execution environments: Local, Docker, SSH, Singularity, Modal, Daytona, Vercel Sandbox. Open source (MIT), easy install (1 command on Linux, macOS, WSL2).
Why a vibe-coder should care
A persistent agent controlled through Telegram is better than a session-based chatbot because it can perform tasks on your behalf even when you’re asleep, grows smarter over time, and is essentially free when not in use. With 242,000 stars in 14 months, it’s one of the most popular self-hosted AI agents out there!
How to install
Copy this and send it to your agent — Claude Code, Codex, any of them:
Install Hermes Agent from https://github.com/NousResearch/hermes-agent following the README — set it up on the current server or connect Modal/Daytona for serverless mode, connect a Telegram bot, and show me the agent's first response. Ask me for API keys if needed.
Runs on a $5 VPS or in serverless mode — nothing to keep running on your laptop. An AI model key required (Claude, OpenAI, or another).
Open on GitHub▌ More finds