Dappnode & Local AI: the perfect match 🚀 With the rise of AI, private and self‑sovereign AI solutions are more necessary than ever. We are building the next generation of Local AI tools based on the principles of Privacy by default, Self‑sovereignty and Censorship resistant 💪 To start with, we’ve built the Dappnode AI Toolkit so you can run modern AI fully on your own hardware 😱

Why Local AI matters? 🤖 Self‑sovereignty: prompts, context, and outputs stay on your machine Privacy by default: local inference via Ollama Censorship resistance: no external API keys for core features Composability: open, modular, and hackable with n8n

What’s inside of our AI toolkit? 🛠️ Ollama: open‑source local LLM runner OpenWebUI: chat interface, “ChatGPT‑like,” but local n8n: automation engine to turn AI into action Immich: private photo/video backup with ML‑powered organization

How does it compare to cloud AI? (ChatGPT, Gemini, Claude...) Dappnode AI toolkit offers: Privacy and control 🔏 No per‑token fees 🆓 Customization ⚙️ Offline resilience 📴 Trade-offs: The newest big models may arrive later ⏳ Performance depends on your hardware 💻 You own maintenance 🔧

How to get started 🔥 Open the Dappstore in your Dappnode UI 💻 Install OpenWebUI, n8n, and Immich 🛠️ Open OpenWebUI package (includes Ollama) 🧠 Pick an LLM 🫵 Chat with it in natural language 🗣️ Open n8n and connect to Ollama 💥 Test a workflow ▶️

Do you have a Dappnode Home? You will be able to run LLMs models in the AI toolkit using your CPU👌 We recommend you to start with smaller models in OpenWebUI (e.g. qwen3:8b) to ensure a decent performance. Note that GPU and unified memory in your hardware significantly improve performance for larger models.

Sounds interesting? 😬 We'll be in @EFDevcon next week in Buenos Aires presenting the AI toolkit and our new AI-ready hardware (spoiler in the first image 🫣) Come to our booth in the World Fair to test it live! See you there! 😉