Creative AI Studio

• 

• 

  Main Screen

• 

Inspiration

The original inspiration came from creating a storyboard app where users can create scenes from images or videos allowing to use multiple AI models and working with multiple assets, with proper asset management and direct prompt based asset and model detection. To be able to simply narrate a story like @jack and @jill went up the hill where the Jack and Jill will automatically be consistent characters and the image / video will consider that when creating the new content. Which then turned out to be a All in one AI content creation suite, using free open source AI models and a workspace to merge them all under one roof, managing and collaborating with multiple assets over multiple models in your browser without any standalone app!

What it does

It allows you to create generative image, video, voice, prompts, and more. using multiple free open source models the timeline feature will support video creation, presentations, etc. Its like Blender and Canva had a baby. The Media viewer will have masking and layers. The Smart Model Selection, Dynamic Response and Result Modals which handle the asset reference, model selection logic, Displaying the results and modifying the parameters of each model are all modular and scalable, will support importing community based plugins, prompts and models with well documented instructions for each.

How we built it

The entire entire application in built on BOLT and BOLT editor only, I strictly challenged myself to build this entirely on the BOLT web app itself, without using external editors like like windsurf or vscode. Many manually edits were made through the browser editor, to avoid AI token usage, or if the edit was too simple.

Challenges we ran into

There were a lot of challenges, deadens where it felt like the AI could no longer fix or implement a particular issue/feature, avoiding excessive token usage, halfway through the project had to purchase tokens again. One particular challenge for me was the timeframe because I got to know about the hackathon very late around 9-10 June I registered the same day and received my pack on 12 June give or take hardly 15 days which also were very challenging to come up with an idea, try BOLT for the first time, solo developer + full time employee.

Accomplishments that we're proud of

Everything that felt like a dead end, completing it felt like a accomplishment, especially when solely relying on the BOLT AI trying to explain it figure out the problem in the console logs, feed it logs or sample code snippets.

What we learned

Working with AI can be extremely tricky, you must have proper goals and plan with certain limits and boundaries or you could get into endless pursuit for new features and countless options, consuming your time and your tokens haha!

What's next for Creative AI Studio

For me this project is in infancy stage there's a lot to be done. to make it the to the vision and goal I have in mind, specially due to the very limited time constrains, some crucial and important core features are still not complete. Export scenes to video. Enhance the Media Viewer to support Out painting, masking images and layers. allow extensive image editing using open-source image editing models. Expand the models and Smart Selection Logic and improvise it with community plugins, prompts, assets etc. Rotated Image Puzzle an easter egg like chrome dinosaur game whenever a request is in progress no one would mind a simple mini game.

Built With

• react
• vite