У нас вы можете посмотреть бесплатно Mindcraft the AI that plays with you Setup Guide - Local Ollama Install Tutorial (2025 Update) или скачать в максимальном доступном качестве, видео которое было загружено на ютуб. Для загрузки выберите вариант из формы ниже:
Если кнопки скачивания не
загрузились
НАЖМИТЕ ЗДЕСЬ или обновите страницу
Если возникают проблемы со скачиванием видео, пожалуйста напишите в поддержку по адресу внизу
страницы.
Спасибо за использование сервиса ClipSaver.ru
Mindcraft AI — Requirements Game + Version Minecraft: Java Edition, supported up to 1.21.6 (recommended: 1.21.6). You’ll run a single-player world and open it to LAN on port 55916. Runtime: Node.js v18+ installed on your machine. Model / API Use one of the following: open windows terminal normally. Local: Install Ollama and pull the default models: ollama pull sweaterdog/andy-4:micro-q8_0 and ollama pull embeddinggemma. Create keys.json from keys.example.json and fill in your chosen provider’s key (skip if using only Ollama). Cloud/API (any single provider is enough): OpenAI, Google Gemini, Anthropic, Replicate, Hugging Face, Groq, Mistral, Qwen, Novita, Cerebras, OpenRouter, xAI, DeepSeek, Hyperbolic, Mercury, or vLLM (self-host). Put the matching API key into keys.json and set your model in the profile (e.g., andy.json). Repository Setup | Clone: git clone https://github.com/mindcraft-bots/min... into the mindcraft folder you created press left shift + right click into the folder you created and press open powershell window it will open the powershell inside of that folder IMPORTANT Install deps: npm install (left shift+ right click) and open a terminal window inside of the folder first. Launch Minecraft, Open to LAN on localhost:55916. Run the bot: node main.js. (Optional) Tasks / Benchmarks If you want the built-in Tasks (automated goals/builds) and evaluations: Install Python and pip (conda optional). pip install -r requirements.txt Run: python tasks/run_task_file.py --task_path=tasks/example_tasks.json (see repo for the full task suite instructions). Config Notes Global toggles live in settings.js (e.g., safety options like code-writing). Per-agent settings (name/model/prompt) live in profiles like andy.json. GitHub Security Caution Code-writing is sandboxed but vulnerable to prompt injection if enabled. It’s off by default; only enable by setting allow_insecure_coding = true in settings.js if you dont understand the risk leave it disabled..