Product Engineer / in progress
VinciUI
Minimal visual editor for composing AI image generation flows with Gemini models.
- Vite
- TypeScript
- TailwindCSS
- Node Editor
- Gemini API
THE PROBLEM
What needed to work better
Image generation experiments often involve repeatedly rebuilding the same prompt, input, and model setup.
THE DECISION
The choice that shaped it
Make a visual node graph the main way to compose a generation flow, while keeping model choice and image feedback close to the work.
HOW IT FITS TOGETHER
A path through the system
- 01Prompt
- 02Image input
- 03Gemini model
- 04Preview
WHAT EXISTS NOW
The result
A browser-based image studio for connecting prompts, inputs, and Gemini models, with session saving and reusable templates.
IN THE BUILD
Details that matter
- Drag-and-drop node editor for prompt chains
- Multiple Gemini model presets with live feedback
- Session-saving and reusable templates
- Lightweight GPU-friendly image preview pipeline