TaskGenie

TaskGenie: Turn Mundane Tasks into Epic Quests with Gemini 3

Comment

Inspiration

We all have endless to-do lists that just add to the pressure. For the "Vibe Coding" theme, we didn't want another boring productivity app. We wanted a tool that matches your energy—a "Vibe Coach" that understands context. We asked ourselves: "What if your task manager cared about your burnout as much as your deadlines?" That’s how TaskGenie was born.

What it does

TaskGenie is a gamified, voice-controlled task manager powered by Gemini 2.5 Flash. It transforms mundane tasks into a quest for productivity:

🛡️ Burnout Shield: The AI analyzes your workload intensity. If you have too many high-priority tasks, it flags "High Burnout Risk" and suggests specific "Vibe Restoration" activities (like "touching grass" or a 15-minute reset). 🗣️ Voice Protocols: You can speak naturally (e.g., "Remind me to ship the feature by Friday night"), and Gemini instantly extracts the structured data (Task, Due Date, Priority) without you typing a thing. 🎮 Gamification: You earn XP for completing tasks, ranking up from "Cadet" to "Galactic Commander," making productivity feel like leveling up in a game.

How we built it

We prioritized speed and "vibes":

AI Engine: We integrated the latest Gemini 2.5 Flash model using the Google Generative AI SDK, solving access issues and boosting performance. We use it for two distinct cognitive tasks: Structured Data Extraction (turning voice to JSON) and Strategic Reasoning (analyzing list health for burnout prevention). Backend: Built on Node.js and Express for a lightweight, fast server. Database: SQLite for reliable local storage that handles our relational data (Tasks, XP, History) without the latency of a cloud DB. Frontend: High-performance Vanilla JavaScript and CSS. We skipped heavy frameworks to keep the interaction snappy and the "vibe" instant.

Challenges we ran into

Model Access & Compatibility: We initially faced access issues with older Gemini models. Debugging the API key permissions led us to upgrade the entire system to Gemini 2.5 Flash, which solved the errors and significantly improved response speed. Balancing Gamification: It was tricky to make the gamification fun without being distracting. We had to iterate on the XP logic (via our migrate_gamification.js scripts) to ensure it rewarded meaningful work, not just spamming easy tasks. Prompt Engineering: Teaching the AI to speak in "Gen Z/Vibe Coding" slang while keeping the JSON output strictly valid required careful tuning of the System Prompt.

Accomplishments that we're proud of

The "Burnout Shield": Watching Gemini correctly identify a stressful Tuesday schedule and intervene with a "Vibe Check" warning feels like magic. It turns the app from a tool into a teammate. Seamless Voice-to-JSON: The voice command feature works incredibly fast thanks to the 2.5 Flash model, creating a "minority report" style experience for adding tasks.

What we learned

AI as a Logic Layer: We learned that LLMs aren't just for chatbots; they are powerful logic engines for internal app state (like determining "Burnout Risk" from raw data). The Power of "Vibe": Building for a specific feeling (fun, futuristic, supportive) made the development process itself more engaging and resulted in a product with a distinct personality. Rapid Iteration: Using migration scripts to evolve our database daily allowed us to pivot quickly based on feedback.

What's next for TaskGenie

Team Vibe: Collaborative task lists where the AI manages team morale. Calendar Sync: Deeply integrating "Vibe Checks" into your actual Google Calendar.

Built With

Share this project:

Updates