Blobby

Inspiration

There are many tasks we have to keep track of, however there are also tons of goals we aim to accomplish and have a struggle balancing with our day-to-day life. Most productivity apps are cold checklists that we abandon after a few days. We wanted something that actually makes us want to come back, so we built a companion that grows with you and cares whether you show up.

What it does

Blobby is a virtual pet that lives inside a full productivity suite. You manage tasks, build habits with streaks, set goals with milestones, run Pomodoro focus sessions, and keep a voice-first journal. As you get things done, your blob feeds, levels up, earns coins, and unlocks a shop of accessories. The more you engage, the more it bonds with you, evolving from a shy, stuttering blob into a hyped-up best friend. An AI chat blob lets you add, edit, and organize everything in plain language, and speech-to-text is built in across the app so you can run your life hands-free. Everything syncs across phone, laptop, and even a Raspberry Pi desk buddy.

How we built it

The backend is a Flask app in Python with a database layer that runs on SQLite locally and Neon serverless PostgreSQL in production, giving each user their own isolated schema for private, cross-device sync. It is deployed on Render with Gunicorn. The AI chat and journal reflections are powered by the Groq API, using tool-calling so the blob can actually create and modify your data. The frontend is vanilla JavaScript with a custom animated SVG blob, and voice input uses the browser Web Speech API. We also ported it to a Raspberry Pi Zero 2 W with a Whisplay HAT for a physical desk companion.

Challenges we ran into

Getting the same SQL to run on both SQLite and PostgreSQL was tricky, since functions like MIN and MAX behave differently and silently broke task completion in production. Making the AI reliably call the right tools, handle multiple actions in one message, and recover when the model returned malformed output took a lot of tuning. Building continuous, voice-first journaling that does not cut off mid-sentence, and keeping the animated blob smooth while syncing real data, were both real engineering problems.

Accomplishments that we're proud of

We shipped a genuinely fun app that is also a complete productivity tool, not a toy. The bonding system, the AI that manages your whole life in plain language, the hands-free voice journaling, and getting it all running live with real cloud sync and even on dedicated hardware are things we are proud of.

What we learned

We learned how to write a portable database layer, how to design reliable AI tool-calling instead of brittle prompts, and how much emotional design matters. Making a productivity app feel like a relationship changed how people actually use it.

What's next for Blobby

Smarter AI that proactively suggests tasks and notices patterns, richer journal insights and mood trends over time, social features so blobs can interact, more shop content and evolutions, and full voice conversations on the Raspberry Pi build.

Built With

• css
• groq
• html
• javascript
• postgresql
• python
• raspberry
• tinyfish