Полное пособие по Hermes Agent: установка и настройка | Все что надо знать за 25 минут!

In this video, I explore Hermes Agent: why I chose it over OpenClaw, the differences between agent-first architecture, how memory, the Learning Loop, skills, and Kanban Swarm work. In the second part, I show how to install Hermes from scratch: choosing a VPS, connecting via SSH, updating the server, creating a Telegram bot, connecting free and cheap neural network providers, launching a web panel, and setting up voice chat via Groq Whisper and Microsoft Edge TTS. All links in Telegram: https://t.me/+XrhmiKgCQdY5MjFi In the video: How an AI agent differs from a regular chatbot; Hermes vs. OpenClaw: memory, architecture, training, and autonomy; Installing Hermes on a VPS; Connecting a Telegram bot; Connecting OpenRouter, GitHub, Allama Cloud, Minimax M3, and Google AI Studio; The Hermes Dashboard; Cron tasks, skills, plugins, Kanban, and memory providers; Voice messaging setup in Telegram via Groq Whisper; Free TTS via Microsoft Edge. All commands, links, and additional materials will be in the Telegram channel. Timeline: 00:00 What you'll get by the end of the video 00:40 Why Hermes Agent and how it differs from OpenClaw 00:59 What is an AI agent 01:29 Learning Loop: How Hermes learns from tasks 02:08 Background curator and GEPA 02:47 Hermes vs. OpenClaw architecture 03:38 Three levels of Hermes memory 04:31 What's improved in Hermes over the past month 04:51 Kanban Swarm and parallel subagents 05:04 Hermes Desktop 05:16 Skill catalog 05:23 Summary of comparison with OpenClaw 06:22 Choosing a VPS server 06:50 Connecting to the server and updating the system 07:23 Installing Hermes 07:52 Running Hermes locally on the server 08:05 Creating a Telegram bot using BotFather 08:23 Receiving User ID and system service installation 08:55 First launch of Hermes in Telegram 09:19 Connecting models via Hermes Setup Model 09:29 OpenRouter: free and cheap models 10:31 Alternatives for replenishment: The benefits of AI and other routers 10:55 Connecting GitHub 11:45 Connecting Allama Cloud 12:43 Connecting Minimax M3 13:40 Google AI Studio and Flash 14:22 Setting up provider switching when limits are reached 14:40 Launching Hermes Dashboard 15:20 Logging in to the web panel via localhost 15:38 Chat, sessions, models, tools, and logs 15:57 Scheduled cron tasks 16:25 Skills, plugins, memory, and communication channels 17:05 Provider keys and system settings 17:23 Documentation, Kanban, and Achievements 17:45 Adding a task to Kanban 18:17 YOLO Mode, Agent Rights, and Tokens 19:36 Setting Up Voice Chat in Telegram 19:54 Groq Whisper for Voice Recognition 20:33 Transferring the API Key and Configuring the Provider 21:29 Restarting Gateway 21:53 Microsoft Edge TTS 22:08 Testing Voice Responses 22:57 What's in the Second Video 23:10 Additional VPS Options 24:08 Finale and Next Step