yimp | Devpost

Inspiration

Creating high-quality advertisements requires expensive tools, technical skills, and time—barriers that small businesses and content creators often struggle to overcome. We wanted to build a platform that democratizes ad creation, making professional-grade video ads accessible to everyone with just an image and a text prompt.

What it does

Our web app uses AI to generate 20-second, high-quality advertisements from simple user inputs. Users upload an image and provide a text prompt, and our platform takes care of the rest by generating a complete ad with AI-generated video footage, emotionally expressive AI voiceovers with natural pauses and lip-syncing, AI-composed background music tailored to the ad’s mood, and professionally structured scripts based on minimal user input. The goal is to eliminate the need for expensive production tools while ensuring that businesses and creators have access to professional-level advertisements effortlessly.

How we built it

We developed the frontend using React and Tailwind CSS, ensuring a sleek, modern, and intuitive user interface. The backend was built with Python and Node.js, allowing smooth processing and integration of various AI models. For music generation, we used Cartesia AI, while Kling AI powered our video generation. Lip-syncing was handled through Sync AI, ensuring that the voiceovers matched the visuals seamlessly. By combining these technologies, we created a streamlined process where a simple user input is transformed into a polished advertisement in just a few steps.

Challenges we ran into

Throughout development, we faced numerous technical challenges. On the UI side, implementing a dynamic background caused unexpected errors, image uploads had inconsistencies, and creating a smooth snap-scrolling experience took longer than anticipated. In terms of music and voice generation, finding reliable APIs that met our needs was difficult, and generating high-quality audio took a significant amount of time, delaying other parts of our workflow. Converting text to video presented its own set of problems, as our initial attempt at browser automation failed, and modal-related issues ended up wasting valuable time. Beyond these, we encountered several overall difficulties, including frustrating merging conflicts and linking errors between the front and backend that required persistent debugging and troubleshooting.

Accomplishments that we're proud of

Despite these challenges, we achieved several major accomplishments. Our UI successfully incorporates a dynamic background, smooth transitions, and a visually appealing, minimalistic design. The AI-generated voiceovers sound incredibly natural, with realistic human emotion and natural pauses, and we developed a system that detects the appropriate tone for the narrator based on the generated script. One of our biggest breakthroughs was creating an algorithm that takes a small user-provided prompt and generates a full professional script, complete with voiceover narration and "movie directions" that ensure optimal video generation. Ultimately, we were able to build a tool that creates high-quality advertisements with ease, making professional video production accessible to anyone.

What we learned

Throughout this project, we learned the importance of choosing the right APIs to optimize performance and reduce delays in the generation process. We also gained a deeper understanding of frontend-backend integration and the challenges that come with managing data flow between different components. Debugging UI animations and improving user experience taught us valuable lessons in web development, while refining our AI-generated content pushed us to explore ways to enhance realism and effectiveness in both voice and video generation.

What's next for Yimp

Moving forward, we plan to improve generation speed to make ad creation even faster and more efficient. We want to enhance video quality and introduce more customization options, allowing users to fine-tune their advertisements to better fit their needs. Expanding language support and offering more voice variations will be a key focus, ensuring accessibility for a broader audience. Additionally, we aim to implement a feature that lets users tweak AI-generated scripts before finalizing their ad, giving them more creative control. Our ultimate goal is to continue refining Yimp so that more creators and businesses can benefit from effortless, high-quality video ad production.