Inspiration:I created DemoScript because I saw many people struggle to present their projects. Even when a project is impressive, its creator may struggle to find the words to explain it. I also wanted to help developers who have difficulty speaking, including those with speech disabilities.
What it does:Users upload a screen recording or record their screen directly in DemoScript. The app analyzes the recording and asks questions to understand the user’s project.It then generates a script divided into scenes that match the recording. Users can record their own voice or use text-to-speech for narration. Once the narration is ready, DemoScript combines the video and audio into a downloadable file.
I used Codex, Kiro CLI, Freebuff, and Antigravity as coding agents. I installed the skills provided by Devpost and used them throughout the project.I used Gemini for video analysis and script writing, and Azure Speech for text-to-speech narration.
Challenges we ran into:I faced several challenges, including rate limits when using Gemini for text-to-speech. Sometimes, the generated script contained short sentences instead of explanations that filled the scene’s duration.
Accomplishments that I'm proud of:I’m proud of completing a whole project with coding agents. This was my first time using skills and my first hackathon.
What we learned:I learned that although we have many ideas, not every project can be built in a week.
What's next for DemoScript:I hope to add built-in audio and video editing.
Guys, the hosted version at demo-script-seven.vercel.app isn’t working the same way as the local version yet. You can run DemoScript locally by following the setup instructions in the GitHub repository and adding your own API keys.
Log in or sign up for Devpost to join the conversation.