šØ Got an idea, a challenge, or just a rough thought?
š Introducing the revolutionary AI CEO!
Just grab your virtual pen and draw your thoughts on a blank canvas. Whether it's a business challenge, a creative idea, or a complex dilemma, the AI CEO will:
āØ Understand your vision
āØ Analyze your problem
āØ Provide detailed, innovative solutions
š No words neededāyour drawing speaks, and AI listens!
Unleash the power of visual problem-solving with the AI CEO. Your ideas, solved effortlessly.
Meet our Team :
Hi, I'm Rajesh Khadka, and together with my skilled video crafter friend Bishesh Marasini , weāre excited to introduce AI CEO ā The Ultimate Problem Solver!
Project Overview :
Sometimes, visuals are the most powerful way to express ideas. You may have a clear thought or a rough idea in mind, but explaining it to others can feel challenging. Thatās where AI CEO steps in!
This innovative application can interpret your drawings, analyze your concepts, and chart out a clear path to address your questions. But thatās not allāAI CEO can tackle a wide range of challenges, from basic to advanced math problems to complex business scenarios. In fact, itās designed to handle virtually any problem you throw at it.
AI CEO is an open-source web application that lets you bring your ideas to life on a canvas. Simply draw your thoughts or describe a problem visually, and the app will process your input and provide solutions directly.
Much like a CEO in a company who tackles diverse challenges, AI CEO is equipped to solve problems across domains, offering clarity and actionable insights for whatever you need. Itās your virtual problem-solving leaderāready to bring your ideas to reality!
Draw it. Express it. Solve itāwith AI CEO!
Core Features of the App :
Extracts the image drawn on the canvas.
Analyzes the canvas image along with user-provided descriptions.
Generates a question based on the drawing and user prompt.
Provides a detailed, step-by-step solution to the generated question.
Some Snaps :
Hero Page :
This is the Landing Page of the application which contains the slogan of the application.
Canvas Page :
This is the page where user draw a image on the canvas. It also contains the sidebar which displays the output after analyzing the image.
Development Process :
1. Ideation:
When the hackathon was announced, I was busy with my mid-semester exams, so I couldn't start right away. After the exams, I had a hard time thinking of an idea. A few days later, after talking with my friend, I decided to create something unique and useful, and that's how the idea for AI CEO came about.
2. Choosing Technologies:
Since this hackathon was organized by Modus, using Modus was a mandatory requirement. In addition, I incorporated the following technologies to build the application:
Google Gemini Generative AI: To process the image, generate a question based on the user prompt and drawing, and feed it into the workflow.
Modus: To generate detailed answers for the questions created by Gemini.
Node.js: For building the server that bridges communication between Gemini and the AI CEO app.
React.js: For developing a responsive and user-friendly frontend.
3. Integrating Modus and Gemini APIs:
As a first-time Modus user, working with its APIs was both exciting and challenging. Integrating Modus and Gemini APIs formed the backbone of the application. Exploring Postman and GraphQL playgrounds during this process added an extra layer of fun and learning.
4. Deployment:
Once the application was built, it was time to deploy:
The frontend was deployed using Netlify, using its live integrations.
The Node.js server was deployed on Render.com after running a few simple commands.
Throughout development, I used the hosted Modus app on Hypermode, which made managing everything at the end much smoother.
5. Presentation:
We wanted to create an engaging and unique presentation video to showcase the app. Since the app is called AI CEO, we decided to simulate a conversation between a top CEO, Elon Musk, and a podcaster named Lex. To make it even more interactive, we utilized AI to generate Elon Muskās voice using a pre-trained model. The entire process was both fun and rewarding, pushing the boundaries of creativity.
And here is the final video made after a multiple take : š
Challenges and How I Overcame Them :
1. Installation Challenges:
I recently switched to Linux (Fedora), so I was still getting used to the new environment. My system wasnāt set up correctly, which caused problems when trying to run the Modus project. After several failed attempts to fix it, I decided to switch back to Windows, where the project ran smoothly.
The images show the Hypermode team members helping me resolve the installation related issue.
2. Image Processing Challenges:
Initially, Modus did not support image processing for generating outputs, making it difficult to work with the canvas-drawn images. To overcome this, I integrated Google Gemini Generative AI to extract possible questions from the images. These questions were then fed into the LAMA model via Modus to generate relevant answers. Additionally, members of the Hypermode team provided valuable assistance, helping me navigate and resolve this challenge effectively.
The images show the Hypermode team members helping me resolve the issue related to image processing.
Some useful Links :
Conclusion :
Overall, the hackathon was an incredible learning experience. I got hands-on with Hypermode and Modus, gained a solid understanding of working with GraphQL APIs, and even discovered some better ways to search on Google. Most importantly, we developed the valuable skill of being patient when our code didnāt work . The support from the community on Discord was invaluable, and it played a key role in shaping the app. The entire experience was amazing, and if you haven't participated in a hackathon yet, I highly recommend giving it a try!
