This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend
What I Built
I built MomTask AI, a simple Hindi/Hinglish AI reminder assistant designed around a real problem my mother could face: turning everyday spoken reminders into organized tasks.
For example:
"Kal subah doodh lana hai"
MomTask understands this as:
- Task: Bring milk
- Date: Tomorrow
- Time: Morning
- Reminder: 08:00 AM
The idea is simple: instead of opening a complicated task manager and typing everything manually, a user can speak or type naturally.
Features
- Hindi/Hinglish task understanding
- Voice input through the microphone
- Add, complete and delete tasks
- Browser reminder notifications
- Custom reminder time
- Local H2 database
- Local AI using Ollama + Qwen3 1.7B
- One-click Windows launcher
Demo
MomTask currently runs locally on Windows.
The project includes a MomTask.bat launcher that starts the required services and opens the application.
After starting it, the app is available at:
text
http://localhost:8080
The microphone button allows Hindi voice input directly from the browser.
Code
GitHub repository:
https://github.com/Shitanshu686/MomTask
The complete source code, README, launcher and Spring Boot project are available in the repository.
How I Built It
The backend is built with Java 17 and Spring Boot.
The basic flow is:
User
↓
MomTask Web UI
↓
Spring Boot Backend
↓
Ollama
↓
Qwen3 1.7B
↓
Structured Task
↓
H2 Database
The AI receives the user's Hindi/Hinglish reminder and extracts:
- task
- date hint
- time
For common Hindi words such as aaj, kal, subah, dopahar, shaam and raat, the application also applies deterministic normalization so the final task data remains predictable.
I also added browser notifications and custom reminder times so the application is useful beyond simply creating a task.
The AI runs locally through Ollama, so the application does not require a cloud AI API key.
Why Does Open Innovation Matter?
Open innovation matters because it makes it possible to experiment, build and adapt technology around real people and real problems.
For MomTask, using a locally running open-weight AI model made the project more personal and privacy-friendly. The AI processing can happen on the user's own computer instead of sending everyday reminders to a remote service.
It also gave me the freedom to experiment with prompts, application logic and the overall user experience while keeping the project simple enough to understand and modify.
What I Learned
The biggest lesson from this project was that an AI application does not need to be complicated to be useful.
The interesting part was not just connecting an AI model. It was combining AI with normal application logic, voice input, reminders, persistence and a simple interface designed around a specific person.
I also learned that local AI performance depends heavily on the available hardware, so designing a lightweight workflow is important.
Built For
I built MomTask AI for my mother as part of the Hacktoberfest 2026 DEV Launch Weekend Build for a Friend challenge.
The goal was to build something useful for a real person rather than making another generic AI demo.
#devchallenge #weekendchallenge #hf26challenge
Top comments (0)