ShrutiBuilder

ShrutiBuilder is an Agentic AI system that turns a simple hackathon idea into a professional website and presentation automatically using Google Gemini.


Inspiration

During hackathons, we often have great ideas but very little time. Most of the time goes into preparing slides, explaining the architecture, and building a basic website for the demo.

Because of this, many teams cannot properly present their ideas.

This inspired us to build ShrutiBuilder — a system that helps participants by automatically creating a website and a presentation from just a simple idea. Our goal was to let developers focus on innovation while AI handles the documentation and presentation work.


What it does

ShrutiBuilder takes a simple input from the user such as:

  • the hackathon problem
  • a short idea description
  • number of slides

After that, the system automatically generates:

  • a complete hackathon presentation
  • a simple project website
  • system architecture explanation
  • workflow description
  • technology suggestions

It works using multiple AI agents that collaborate together to produce the final result.


How we built it

We built ShrutiBuilder using modern AI tools and frameworks.

Main technologies used:

  • Google Gemini API for AI reasoning and content generation
  • Python for backend development
  • LangChain for managing prompts and AI workflows
  • LangGraph for coordinating multiple AI agents
  • python-pptx for automatically creating PowerPoint presentations

Different AI agents handle different tasks such as understanding the idea, planning slides, writing content, and generating the final presentation.


Challenges we ran into

While building this project, we faced several challenges.

One challenge was making sure that different AI agents could work together smoothly.

Another challenge was generating well-structured presentation slides automatically.

We also had to ensure that the final output looked professional and clear enough for a hackathon presentation.

Managing the workflow between multiple agents required careful design.


Accomplishments that we're proud of

We are proud that we were able to build a system that can automatically generate both a website and a presentation from just an idea.

The system demonstrates how Agentic AI can work like a small virtual team, where different agents handle different responsibilities.

It shows the potential of AI to simplify complex tasks for developers and innovators.


What we learned

Through this project, we learned many important concepts such as:

  • building multi-agent AI systems
  • working with Google Gemini APIs
  • designing AI workflows using LangGraph
  • prompt engineering for better AI responses
  • structuring automated presentations

This project also helped us understand how AI can automate creative and technical workflows.


What's next for ShrutiBuilder

In the future, we want to improve ShrutiBuilder by adding more powerful features.

Some of the ideas we plan to explore include:

  • generating fully deployable project code
  • creating interactive architecture diagrams
  • adding voice input for project ideas
  • automatically generating GitHub repositories
  • supporting more types of presentations and demos

Our long-term goal is to turn ShrutiBuilder into a complete AI assistant for building and presenting innovative ideas.


ShrutiBuilder – Your AI Hackathon Partner.

Built With

Share this project:

Updates

Submission history