Build private, local AI desktop apps . - Offline AI Boilerplate

Building a 100% offline, private AI desktop app is a configuration nightmare. Getting Electron IPC bridges, Vite, LanceDB native modules, and local LLMs to compile together usually takes days of frustrating debugging. I built this production-ready boilerplate so you can skip the 40-hour setup headache and jump straight into building your own AI tools.

Intercom — Startups get 90% off Intercom + 1 year of Fin AI Agent free

Startups get 90% off Intercom + 1 year of Fin AI Agent free

Hi Product Hunt! 👋 When I was building my last app (Unified Gallery), I dealt with a lot of cloud integrations. But for AI, sending private notes, personal journals, or proprietary code to public cloud APIs is a massive privacy risk. I wanted to build a 100% offline, air-gapped AI desktop app. But getting Electron IPC bridges, Vite bundlers, LanceDB native C++ modules, and node-llama-cpp to all compile together was a 40-hour configuration nightmare. I finally got a stable architecture working perfectly. Instead of letting that code sit on my hard drive, I packaged the entire RAG pipeline into a clean boilerplate. It includes: ⚡ React + Vite frontend 🖥️ Electron backend (Node.js) 🗄️ LanceDB Vector Search (pre-configured) 🧠 Llama.cpp engine for local .gguf models 📄 PDF document chunking ready to go If you want to build private AI tools without fighting native module errors for a week, you can grab the source code today.

Intercom — Startups get 90% off Intercom + 1 year of Fin AI Agent free

Startups get 90% off Intercom + 1 year of Fin AI Agent free