Summary
Matthew Berman walks through a complete setup of Hermes, an AI agent platform built by Nouse Research, demonstrating how quickly it can be deployed via Hostinger’s managed hosting service. The video covers the full onboarding flow — connecting an OpenAI API key, selecting a workspace, and launching a working Hermes instance in under two minutes — then dives into what distinguishes Hermes from similar tools like OpenClaw.
Berman highlights Hermes’s built-in skill library, which ships pre-configured with capabilities including a Claude Code skill, an Excalidraw diagram skill, and an Asky video tool. He shows how to install additional skills directly from GitHub by copying a raw markdown file into the Hermes interface — and demonstrates the platform’s self-healing behavior when a skill’s required dependency file is missing. The video also covers scheduled task automation (e.g., daily calendar briefings), persistent memory that builds a user profile over time, agent soul documents analogous to OpenClaw’s identity.md, and profile-based agent segmentation for separating marketing versus development workflows.
On the model side, Hermes supports a wide range of providers out of the box — OpenAI (including GPT-5.4 Mini), Anthropic, Deepseek, Gemini, Copilot, and local inference via LM Studio. The video is well-suited for users exploring personal AI agent setups who want a managed, low-configuration alternative to self-hosted frameworks.
📺 Source: Matthew Berman · Published June 28, 2026
🏷️ Format: Tutorial Demo
