Prompt2Project

Inspiration

Coming up with the right project idea is hard, especially under tight time constraints. While brainstorming, we reflected on how many ideas we had to evaluate and how difficult it was to determine what was both feasible and impactful within our constraints. That very challenge became our motivation: to build a tool that helps others solve the same problem. We wanted to generate project ideas based on a team's existing technical skills, time constraints, and project requirements. We also wanted to provide a clear trajectory of tasks and responsibilities to help bring a project to life.

What it does

Prompt2Project makes it easy to generate ideas and delegate tasks to accelerate productivity. First, you input your team members, their technical skills, and experience levels. Then you define your project requirements and preferred difficulty level, and the AI suggests tailored project ideas. Once you select an idea you’re satisfied with, the tool generates a detailed breakdown of tasks assigned to each team member. These tasks can be edited as needed, and users can check them off while tracking progress through a progress bar.

How we built it

We built Prompt2Project using Python and Streamlit to create the forms and web pages. We used Gemma 4 to generate project ideas based on user-defined constraints and Gemma 3 to define the tasks. The models were used to parse and structure the output into a formatted response suitable for the application.

Challenges we ran into

This was our first time using Streamlit, so there was a learning curve in understanding session state management and navigating between pages. While Streamlit provides useful built-in components like forms, we found it limited in terms of styling and customization. It took time to refine the UI to a point we were happy with. We also encountered issues with the Gemini model not consistently producing parsable output, which required experimenting with different models.

Accomplishments that we're proud of

We are proud of successfully using Streamlit, a framework that was new to us. We’re also proud of our prompt engineering work and successfully leveraging Gemini models to generate the structured outputs we needed. Additionally, we built a user interface that makes it easy to generate ideas and manage tasks effectively.

What we learned

During this hackathon, we learned how to collaborate more effectively as a team. We were fortunate not to run into merge conflicts, and we maintained clear communication throughout the project. We also learned how to accelerate development using AI tools like Gemini and Claude to lay the groundwork, so that we can focus on the finer details.

What's next for Prompt2Project

Next, we plan to improve the UI/UX experience, potentially using Figma to design a more polished and customized interface. We also want to add more collaborative features to help teams stay aligned and track progress more effectively.

Built With

• gemini-api
• python
• streamlit