Summary
Keith AI delivers a 34-minute comprehensive walkthrough of Hermes Desktop, the graphical control application for the Hermes AI agent platform by Nautilus Research (hermesagent.nautilusresearch.com). The video explains why the creator migrated from managing Hermes through Telegram — which became unmanageable with multiple simultaneous agents and hundreds of pages of chat history — to a dedicated desktop interface with proper session management and file organization.
The tutorial covers the full setup stack: installing the desktop app in under three minutes, connecting to LLM providers including OpenAI’s cloud API and local models via compatible inference servers, deploying Hermes on a Hostinger VPS using Docker, and establishing a Tailscale-based remote gateway so the same persistent Hermes instance can be accessed from any device. Specific steps include finding the open container port, retrieving the Tailscale IP address, and entering the combined address in the Hermes gateway settings.
Beyond basic setup, the video explores the new Profiles feature — which stores distinct configurations pairing different LLM providers with different system prompts for different task types — alongside Artifacts for searching across images, files, and links shared in previous sessions, Cron Blueprints for scheduling automated tasks, and sub-agent configuration. The video also shows how Hermes Desktop consolidates conversations from connected Telegram and WhatsApp channels into a single, searchable interface, resolving the multi-channel chaos that originally motivated the switch.
📺 Source: Keith AI · Published June 19, 2026
🏷️ Format: Tutorial Demo
