-
Live demo of TaskMind converting raw meeting notes into structured JSON tasks using Gemini AI
Inspiration
In most teams, meeting notes are messy, unstructured, and hard to follow. Important action items often get missed, owners are unclear, and follow-ups become confusing. I experienced this problem repeatedly in group discussions and online meetings, which inspired me to build a simple AI-powered solution.
What it does
TaskMind converts raw meeting notes into clean, structured, and actionable tasks.
From a block of unorganized text, it automatically extracts:
- Task description
- Task owner (if mentioned)
- Deadline (if mentioned)
The output is returned instantly as structured JSON, making it easy to track and use in other tools.
How I built it
The application is built using Python and the Gemini 2.5 Flash API.
Users paste their meeting notes into a simple Gradio-based web interface.
The AI processes the text using prompt engineering and returns structured task data in JSON format.
The project was developed and tested entirely in Google Colab, making it lightweight and easy to demo.
Challenges I ran into
One of the main challenges was handling unstructured and incomplete inputs, such as missing owners or deadlines. Another challenge was ensuring consistent JSON output from the AI while keeping response time fast and reliable.
What I learned
This project helped me understand how to design effective AI prompts, work with generative AI APIs, and build quick, user-friendly demos for hackathons. I also learned how to balance simplicity and functionality when building under time constraints.
What's next for TaskMind
Future improvements could include task prioritization, exporting tasks to productivity tools, and supporting multiple languages.
Log in or sign up for Devpost to join the conversation.