Haru
This project was initiated to solve specific technical and domain challenges. Integrating multiple local AI models and productivity tools into a cohesive cross‑platform desktop app without relying on cloud services.
About This Project
Integrating multiple local AI models and productivity tools into a cohesive cross‑platform desktop app without relying on cloud services.
Architected a modular Rust backend, integrated gguf LLMs, Whisper and Piper for voice and language, used SolidJS for UI, built dynamic courses and interactive 3D visualisations, all running locally.
Delivered a privacy‑first personal study assistant that merges multiple AI capabilities and runs entirely offline.
AI & Full-stack Developer
2025
Public
Personal
Technology Stack
Project Story
Integrating multiple local AI models and productivity tools into a cohesive cross‑platform desktop app without relying on cloud services.
Architected a modular Rust backend, integrated gguf LLMs, Whisper and Piper for voice and language, used SolidJS for UI, built dynamic courses and interactive 3D visualisations, all running locally.
Delivered a privacy‑first personal study assistant that merges multiple AI capabilities and runs entirely offline.
Insights & Takeaways
Highlights
- Case study content natively baked into the project dataset.
- Clear storytelling built around the specific problems faced and the technologies used.
Challenges
- Strict focus on performance and maintainability.
- Selecting standard tools to ensure scalability: Tauri, Rust, SolidJS, Python, Whisper, Piper, gguf LLMs, Three.js
- Architected a modular Rust backend, integrated gguf LLMs, Whisper and Piper for voice and language, used SolidJS for UI, built dynamic courses and interactive 3D visualisations, all running locally.
Lessons Learned
- Delivered a privacy‑first personal study assistant that merges multiple AI capabilities and runs entirely offline.
