Whisper Board

• 

  whisperboard

Inspiration

I always have clear cinematic visions in my head, but I seriously cannot draw past basic stick figures. Trying to sketch out storyboards was a massive creative bottleneck for me. I wanted a tool where I could just use my voice live to instantly turn my ideas into sketches without needing traditional art skills.

What it does WhisperBoard is a voice-powered multi-scene storyboard generator. Creators can click record, speak a scene description (like "aerial shot of a city"), and the app automatically transcribes it and renders a professional storyboard sketch live on a digital director's pad UI.

How i built it Backend: FastAPI handling async routes and server logic.

AI & Voice: OpenAI Whisper API for live audio transcription and prompt processing.

Storage: Backblaze B2 cloud storage for managing media assets.

Frontend: A clean, custom web interface styled like a classic storyboard pad.

Challenges we ran into Debugging terminal paths, getting the Python virtual environment and genblaze_core packages to link up properly, setting up Backblaze B2 bucket regions (us-east-005), and syncing live microphone inputs with real-time backend rendering without breaking the flow.

Accomplishments that we're proud of Pulling off a fully functional, end-to-end generative AI app that works completely hands-free using just voice prompts, and getting it fully built and deployed for my third hackathon.

What we learned How to wire up FastAPI backends with live audio streams, manage API credentials and cloud regions smoothly, and push through terminal debugging roadblocks under pressure.

What's next for WhisperBoard Adding options for different art styles, exporting whole storyboard sequences directly into video editing tools, and improving multi-scene continuity.

Built With

• backblaze-b2
• css
• fastapi
• html
• javascript
• openai
• python
• uvicorn
• whisper