Clip Oracle was created to solve a problem many creators face: turning long videos into short, shareable content takes too much time. Podcasters, streamers, educators, and social-media teams often record hours of content, then need to manually watch it again to find the best moments. We wanted to make that process faster and easier.
Our platform allows users to upload a long-form video, such as a podcast episode, interview, webinar, livestream, or tutorial. Once the video is uploaded, Clip Oracle processes the content and analyzes its transcript using AI. The system looks for moments that are interesting, useful, funny, emotional, surprising, or likely to hold an audience’s attention. It then suggests short clips with timestamps, descriptions, and confidence scores, helping users decide which parts of their video are most worth sharing.
We designed Clip Oracle with creators in mind. The user experience is simple: upload a video, wait for it to process, review the suggested clips, and choose the moments to repurpose for platforms like TikTok, Instagram Reels, YouTube Shorts, LinkedIn, and other social channels. The goal is not to replace a creator’s judgment, but to remove the most time-consuming part of the editing process: searching through long recordings.
To build the product, we used AWS services to create a scalable workflow. Amazon S3 stores uploaded video files and processed media. AWS Lambda handles background tasks after upload, allowing the platform to process content without requiring users to stay on one screen. Amazon Bedrock provides the AI capabilities used to analyze transcripts, identify strong moments, generate descriptions, and rank potential clips. Amazon DynamoDB stores project details, processing status, transcript references, and clip recommendations.
We used API Gateway to connect the frontend with the backend services, allowing the application to send requests, check processing progress, and display AI-generated results. Amazon CloudWatch supported testing and debugging by helping us monitor logs and identify issues across the workflow. AWS CDK helped us define the infrastructure as code, making it easier to set up, update, and manage the project as a team.
The project was developed collaboratively. We led the product direction, UI and UX design, AI prompt development, AWS setup, testing, and pitching. Backend development focused on connecting the application to AWS services, managing the upload and processing workflow, and ensuring the AI results could be returned reliably to users. Frontend development turned the designs into a working interface, including the upload flow, project dashboard, processing status, and clip recommendation views. We also created customer-journey stories and test cases based on realistic creator scenarios to validate that the experience was clear and the suggested clips were useful.
Clip Oracle shows how AI and cloud services can help creators get more value from the content they have already created. Instead of spending hours searching through a recording, users can quickly discover the moments with the strongest potential and spend more time creating, publishing, and connecting with their audience.
Loading article