butlrapp

Butlr is an AI task manager you talk to like a human assistant. It understands intent, handles complex dates, tracks time, and organizes tasks automatically no menus, no friction.

Comment 2

Inspiration

The inspiration for Butlr came from a simple observation: most task managers feel like work. Traditional apps require tedious manual entry—clicking through date pickers, toggling priority flags, and typing out descriptions in rigid fields. We wanted to build something that felt less like a database and more like a conversation. Inspired by the simplicity of the ChatGPT mobile app and the elegance of high-end personal assistants, we set out to create a "Butlr" for the digital age an assistant that simply listens and acts.

What it does

Butlr turns raw natural language into structured, actionable tasks. You can say: "Hey Butlr, remind me to check the oven in 20 minutes" or "Every Monday morning at 9 AM, I need to sync with the team."

  • Conversational Intelligence: Processes complex relative timeframes and intent .
  • Visual Schedule Management: Offers a smart, grouped list view and a high-fidelity timeline view.
  • Live Context: Features a dynamic Current Time Indicator so you always know your position relative to your goals.
  • Proactive Organization: Automatically calculates priorities based on task sentiment and keywords.

How we built it

We chose a modern, high-performance stack to ensure a premium user experience:

  • Frontend: Flutter allowed us to build a beautifully animated, glassmorphic UI that runs flawlessly on Web and Mobile.
  • Backend: Serverpod served as our "missing server for Flutter," providing a robust ORM and seamless communication between the client and our PostgreSQL database.
  • AI Brain: We integrated Llama 3.1-8B via Novita AI. We implemented a custom JSON-mode prompt wrapper that ensures the model returns strictly valid schema.

Challenges we ran into

One of the biggest hurdles was Temporal Logic. Parsing "next Tuesday" or "bi-weekly" requires the AI to have accurate state consciousness. We had to carefully inject the current server time and locale into every prompt to ensure the LLM didn't hallucinate dates. Additionally, deploying a Flutter web app within a Serverpod environment on Serverpod Cloud required precise pathing configuration for the base-href and asset routing.

Accomplishments that we're proud of

We are incredibly proud of the UI/UX fidelity. Getting the "glass" look right while maintaining 60fps performance during swipes and transitions was a challenge. More importantly, we successfully deployed a full-stack Dart application to a production cloud environment, bridging the gap between a local prototype and a live, shareable product for the hackathon.

What we learned

Building Butlr taught us the power of Type-Safe Full-stack development. Using Dart for both the server and client (via Serverpod) significantly reduced the friction of API development. We also gained deep insights into prompt engineering for JSON extraction and the nuances of handling recurring event projections in a relational database.

What's next for butlrapp

The journey doesn't end here. The next phase for Butlr includes:

  • True Voice-First Integration: Using Whisper or similar STT models for hands-free task management while driving or cooking.
  • Smart Recommendations: Using historical data to suggest optimal times for tasks based on the user's energy levels.
  • Collaborative Butlr: Shared "Households" or "Team" sessions where Butlr can coordinate tasks between multiple people in a single chat.

Built With

Share this project:

Updates