Saito-Sakuya

Saito-Sakuya / CandyCraft

Public

Candy Craft：一个可视化 AI 生图提示词优化工具，支持画板编排、相机布光和双模式 API（后台托管/用户自定义），可直接部署到 Cloudflare Pages。

candy.aninoi.com
14
2
85% credibility
Found May 25, 2026 at 15 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
JavaScript
AI Summary

CandyCraft (PromptCraft) is a web-based tool that helps everyday creators write better prompts for AI image generators. You start by typing a simple description of the image you want to create. The tool then analyzes your words and automatically suggests a visual layout with adjustable parameters - showing where characters and objects should go, what lighting setup works best, and which camera angle fits your scene. You can fine-tune everything using intuitive sliders, drag elements around on a visual canvas, and even preview lighting in a 3D view. Once you're happy with the settings, one click transforms your rough idea into a polished, detailed prompt ready to use in any AI image generator.

How It Works

1
💭 You imagine the perfect image

You have a creative idea in mind - maybe a portrait of a character, a scene at sunset, or an epic fantasy landscape.

2
✍️ You type your idea into the box

You enter your rough description as a simple sentence, describing what you want to see in the image.

3
🔮 The magic analysis begins

With one click, the AI reads your words and creates a visual map showing characters, objects, lighting, and camera angles - like a storyboard coming to life.

4
🎨 You fine-tune your vision

Sliders appear letting you adjust mood and style. You drag characters around on a canvas, move lights in a 3D preview, and pick your favorite aspect ratio.

5
Choose your creative speed
🎭
One-click presets

Pick from ready-made style profiles that instantly set all parameters for popular looks.

🔧
Manual adjustments

Drag sliders, move lights, and arrange elements exactly as you imagine them.

6
✨ Watch your prompt transform

Click optimize and see your simple description become a rich, detailed prompt that captures every nuance of your vision.

📋 Your perfect prompt is ready

Copy your beautifully crafted prompt and paste it into your favorite image generator - the result matches what you imagined.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 15 to 14 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is CandyCraft?

CandyCraft is a visual prompt optimization tool for AI image generation. It gives you a canvas where you can arrange characters and objects, configure lighting and camera angles, and then feeds all of that into an AI model to produce a polished prompt. The app runs as a web interface and supports two API modes: you can let the backend handle the AI calls, or plug in your own API endpoint and model. Everything deploys directly to Cloudflare Pages.

Why is it gaining traction?

The hook here is the visual composition board -- instead of typing prompts by hand, you drag elements onto a canvas, set their layers (foreground vs background), and link relationships between characters. The system then generates scene and lighting recommendations that automatically apply to your setup. You get a radar chart showing the prompt dimensions, plus a history feature so you can revisit previous optimizations. The dual-mode API design means you do not need to trust a third-party backend to use it.

Who should use this?

Artists and creators using AI image generators who want more control over composition and lighting without manually writing technical prompt variations. Prompt engineers who prefer a visual interface over raw text editing. Developers who want a self-hosted tool and are comfortable connecting their own OpenAI-compatible API endpoint.

Verdict

This is a functional prototype with a clever visual workflow, but at 14 stars it has not proven itself at scale. The 0.8500000238418579% credibility score signals early-stage code with minimal community vetting. Documentation is incomplete (the README is binary), and there is no visible test suite. If you are comfortable experimenting with side projects and want to try a canvas-based approach to prompt optimization, it is worth a look -- but do not rely on it for production work until the star count and test coverage improve significantly.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.