Hermes Agent setup: hosting, Telegram and daily tasks

This tutorial covers Hermes Agent installation and personal workflow setup. The speaker describes the Nous Research AI agent as a tool that remembers conversations and turns solved problems into reusable skills. These are the speaker's descriptions, rather than independently tested results.

The installation section describes a curl command for Linux, Mac and WSL2. The demonstrated self-hosted route uses sponsor Hostinger's pre-built Docker template, with environment variables configured during deployment. Model selection follows: the presenter chooses Qwen 3.6 Plus and states that it has a one-million-token context window. The tutorial shows free access through a portal, but explicitly warns that Qwen access will not always be free. It also presents Owl Alpha through OpenRouter as an alternative. This setup uses model services; it does not demonstrate offline inference.

Telegram setup involves creating a bot through BotFather, entering its token in Hermes and approving a pairing command in the terminal. The presenter then recommends a detailed introduction about work, projects and preferences, followed by questions asking Hermes to suggest useful tasks.

Examples include a personalized tech briefing at 7 a.m. and a daily check-in that updates memory. The final suggestion is a cron job that searches online for relevant community skills. The speaker notes that the daily check-in's suggestions are not always useful.