Artloop : Sketch-to-Art platform

Ever had an idea in your head, sketched it out, and thought: "Yeah... this looks awful"?

Same.
That’s why I built ArtLoop – a sketch-to-art platform that turns your messy hand-drawn ideas into polished, styled illustrations using AI and computer vision.

Just draw something rough, describe what you're aiming for, and let ArtLoop do the magic.

What Makes ArtLoop Special?

1.Draw Freely – Clean canvas with brush, pencil, eraser, custom colors and sizes
2.AI That Gets It – Combines Google Cloud Vision with Flux.1 Schnell model to truly understand your sketch
3.Style It Your Way – Realistic, Anime, Cartoon, Oil Paint, Cyberpunk, Sketch, and more
4.Smarter Prompts – Your sketch gets analyzed (colors, objects, context) to enhance your prompt
5.Credit-Based – You start with 6 free generations
6.Secure Auth – Appwrite, JWT, and hashed passwords for a smooth login experience

How ArtLoop Works?

1. Sketch Capture
   The sketch is converted to a base64 string in the frontend.

2. Backend Processing
   The base64 string is sent to the backend and converted into a binary buffer.

3. Vision Analysis
   The buffer is sent to Google Cloud Vision API, which analyzes objects, text, colors and visually similar images

4. Prompt Assembly
   When the user adds a prompt and selects a style, I combine as :

   Vision API results + User prompt + Style preset

5. Image Generation
   This combined prompt is sent to FLUX.1-schnell model, which returns a final image link.

Doodle something random. Write a weird prompt. Let us cook...

Let ArtLoop do the rest, just pure creativity (or chaos, but hopefully not)

Project Live here : Artloop

Github links - Frontend | Backend