<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
    <channel>
        <title>Gemma 4 LOCAL : Remplace Claude GRATUITEMENT (100% offline 🤯)</title>
        <link>https://tube-cycle-3.apps.education.fr/videos/watch/a549b108-1383-4517-9044-bd72392104b8</link>
        <description>Ton propre LLM local, chez toi, 100% offline, zéro abonnement, zéro fuite de données. Google DeepMind vient de sortir Gemma 4. Un modèle open source qui rivalise avec Claude Sonnet et dans cette vidéo, je t'explique comment l'installer sur ton Mac ou ton PC, le configurer avec LM Studio, activer les connecteurs MCP via Docker, et le connecter à tous tes outils (n8n, OpenClaw, Claude Code). À la fin, tu as ton propre LLM local. 🎁 Guides pratiques à télécharger gratuitement : https://tally.so/r/PdBz2P 👉 Ce que vous allez apprendre : ✅ Pourquoi installer un LLM local (économies, vie privée, indépendance) ✅ Ce qu'est Gemma 4 et pourquoi c'est une révolution open source ✅ Quelle machine / quel modèle selon ton budget (Mac Apple Silicon + Windows NVIDIA) ✅ Comprendre la quantisation : Q4, Q8, BF16 — lequel choisir ✅ Installer et configurer LM Studio pas à pas ✅ Activer les connecteurs MCP via Docker dans LM Studio ✅ Connecter ton LLM local à n8n, OpenClaw et Claude Code via API 🛠 Outils et ressources mentionnés : 🔗 LM Studio → https://lmstudio.ai 🔗 Gemma 4 (Google DeepMind) → https://ai.google.dev/gemma 🔗 Cloudflare Tunnel → https://developers.cloudflare.com/cloudflare-one/connections/connect-networks/ 🔗 Docker → https://www.docker.com 🔗 n8n → https://n8n.io 🔗 OpenClaw → https://www.hostinger.fr/av10 🔗 Claude Code → https://claude.ai/code MES PROGRAMMES DE FORMATION (15 mois) : 🚀 NoCode Bootcamp – Deviens chef de projet NoCode et construis des apps sans coder, en alternance → https://bit.ly/4jlh1Rr MES PROGRAMMES DE FORMATION (2 mois) : 🤖 Formation Make – Automatise tes process avec Make → https://bit.ly/3EeaDfF 💻 Formation Bubble – Crée ton SaaS sans coder avec Bubble → https://bit.ly/43Dm2A7 🎨 Formation Webflow – Crée un site pro avec Webflow → https://bit.ly/42glYUy MES AUTRES COMPTES : 🌍 Mon site web – https://www.illith.com/ 👨‍💻 LinkedIn – https://www.linkedin.com/in/antoine-vigneau33/ 📱 TikTok – tiktok.com/@illithformation 📸 Instagram – https://www.instagram.com/illith_bordeaux/ ⌚️ Les temps forts de la vidéo : 0:00 Introduction 0:44 Ce qu'on va voir dans cette vidéo 1:12 C'est quoi un LLM ? 1:31 Pourquoi installer un LLM en local ? (4 raisons) 3:33 Le hardware : pourquoi c'était pas possible avant 4:09 Mac vs Nvidia — quel matériel choisir ? 4:28 Mac Mini comme serveur IA — le calcul qui change tout 5:17 Gemma 4 — le nouveau modèle open source de Google 7:22 Quelle machine pour quel modèle ? (tableaux Mac + Windows) 9:30 Q4, Q8, BF16 — la quantisation expliquée simplement 10:12 Ollama vs LM Studio vs Jan — comparatif des solutions 11:37 Installation de LM Studio (pas à pas) 12:23 Configuration générale &amp; mode développeur 14:04 Télécharger Gemma 4 dans LM Studio 16:46 Configurer le modèle (contexte, température, système) 19:44 Premier test en direct avec Gemma 4 20:24 Connecter des outils MCP à votre LLM 22:00 Docker + MCP Toolkit — le catalogue complet 24:48 Demo live — MCP YouTube Transcript en action 27:29 LM Link — partager votre modèle avec d'autres machines 28:39 Connecter votre LLM via API (N8N, Make, etc.) 31:44 Sécuriser vos appels API avec un token 34:26 Problème localhost → solution tunnel Cloudflare 36:18 Connecter N8N avec l'URL tunnel 37:05 Créer un agent N8N avec votre modèle local 39:09 Configurer Claude Code avec LM Studio 42:01 Connecter Open Claw à votre modèle local 45:45 Conclusion — limites actuelles &amp; ce qui arrive 46:02 Outro</description>
        <lastBuildDate>Tue, 05 May 2026 17:40:20 GMT</lastBuildDate>
        <docs>https://validator.w3.org/feed/docs/rss2.html</docs>
        <generator>PeerTube - https://tube-cycle-3.apps.education.fr</generator>
        <image>
            <title>Gemma 4 LOCAL : Remplace Claude GRATUITEMENT (100% offline 🤯)</title>
            <url>https://tube-cycle-3.apps.education.fr/client/assets/images/icons/icon-1500x1500.png</url>
            <link>https://tube-cycle-3.apps.education.fr/videos/watch/a549b108-1383-4517-9044-bd72392104b8</link>
        </image>
        <copyright>All rights reserved, unless otherwise specified in the terms specified at https://tube-cycle-3.apps.education.fr/about and potential licenses granted by each content's rightholder.</copyright>
        <atom:link href="https://tube-cycle-3.apps.education.fr/feeds/video-comments.xml?videoId=a549b108-1383-4517-9044-bd72392104b8" rel="self" type="application/rss+xml"/>
    </channel>
</rss>