2026 · Personal Project · Hardware × AI
The Camera for Creators
AI is being used redundantly — treated as post-production instead of embedded at capture. Voice is the most human part of art direction, yet execution waits until post.
Directors' creative decisions happen naturally on-set through voice — 'make it rain', 'bring up those highlights', 'make the flowers look more computer generated'. These are real-time creative calls happening in natural language, but execution gets deferred for hours, sometimes days.
A new kind of camera — one that listens on-set, no screen needed
Voice in. Image out. Real-time generative editing at the moment of capture. Costly post-production workflows collapse into 4-pence, 5-second voice prompts.
Visual UI magic enhances perceived latency. Edge deployment handles actual latency. The Gemma model runs on-device — only cloud round-trips for generative edits with Nano Banana. Voice metadata trains the AI to become the creator's conversational collaborator.
Industrial design in Fusion 360 — from voice-first concept to physical form.
Build process
At the moment of capture.
Year
2026
Role
Independent
Field
Hardware × AI Product
Built with
Raspberry Pi 5, Gemma 4 E2B, LiteRT, Fusion 360
Next project
Next project
Next project
