PersonalPublic2025

Haru

Driven by a personal need to study more efficiently, stay deeply motivated, and eliminate time-wasting distractions, I developed Haru. It is a comprehensive platform that houses everything in one place, seamlessly integrating high-level AI models like FastText without relying on cloud services.

Haru

About This Project

Creating an all-in-one personalized environment to eliminate time-wasting, boost study motivation, and manage high-level AI binaries locally.

Engineered a complex architecture that orchestrates custom-compiled binaries (such as FFmpeg, Llama.cpp, and FastText) running via WSL on Windows, unifying all study resources into a single application.

Transformed my daily workflow by centralizing study resources and AI assistance, drastically improving efficiency, focus, and motivation to achieve more.

Role

AI & Full-stack Developer

Year

2025

Status

Public

Type

Personal

Technology Stack

TauriRustSolidJSPythonWSLFastTextLlama.cppFFmpeg

Project Story

The Challenge

Creating an all-in-one personalized environment to eliminate time-wasting, boost study motivation, and manage high-level AI binaries locally.

The Approach

Engineered a complex architecture that orchestrates custom-compiled binaries (such as FFmpeg, Llama.cpp, and FastText) running via WSL on Windows, unifying all study resources into a single application.

The Outcome

Transformed my daily workflow by centralizing study resources and AI assistance, drastically improving efficiency, focus, and motivation to achieve more.

Insights & Takeaways

Highlights

  • Integrated advanced AI models like FastText and Llama.cpp for intelligent, localized insights.
  • Unified a fragmented study workflow into one centralized, distraction-free productivity platform.

Challenges

  • Seamlessly orchestrating and communicating with multiple separately compiled WSL binaries from a Windows desktop interface.
  • Ensuring high responsiveness while managing heavy background ML processes.

Lessons Learned

  • Mastered complex cross-platform execution environments and process management.
  • Proved that targeted, personalized tooling heavily outperforms generic productivity apps for maintaining long-term motivation.

Related Work