> ## Content Index
> Fetch the complete content index at: https://thibaultmilan.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Super Fiesta
- URL: https://thibaultmilan.com/projects/super-fiesta/
- Published: 2025-08-06T12:00:00.000Z
- Updated: 2025-09-22T11:13:02.000Z
- Description: A Python tool that transforms party photos into perfectly square, Instagram-ready images with professional watermarking and intelligent batch processing. Built with AI assistance for just $5, it replaces heavyweight GUI tools like Photoshop macros and Squash with a lightweight CLI that handles resiz
- Author: Thibault Milan
- Tags: project, #nav, Python, PIL/Pillow, Claude Code, AI-Assisted Coding, Photography, #Import 2026-06-21 19:35

## The Challenge

Every party left me with hundreds of pictures that were too heavy, not square, unbranded, and badly named. I spent more time juggling Photoshop actions and Squash than enjoying the memories. My tiny Python script—hacked together in an afternoon and inspired by [@roast.london.roast](https://www.instagram.com/roast.london.roast/?hl=fr&ref=thibaultmilan.com)'s clean aesthetic—only did basic resize and... that was it.

## What Super Fiesta Does

- **Smart Resizing** to 1080×1080px squares, centers photos, fills backgrounds
- **Batch Processing** entire folders with optional macOS tag filtering
- **Adjustable Quality** JPEG compression settings
- **Professional Watermarking** \- PNG overlay with size, position, opacity, and margin controls
- **Intelligent Renaming** \- Original, sequential, or Event-Date patterns with conflict detection
- **Persistent Configuration** via `sf_config.ini`
- **User-Friendly CLI** with input validation and progress tracking

## AI-Assisted Development

I'm not a full-time Python dev; I wanted to move fast, stay clean, and learn along the way. Claude Code bills per token, so with focused prompts I:

- **Rewrote the README** to be genuinely onboarding-friendly
- **Hardened the script** with proper error handling, RGB/quality checks, and context managers
- **Implemented watermarking** end-to-end with flexible positioning
- **Added pattern-based renaming** with preview and conflict detection

**Total cost: \~5,800 tokens ⇒ \~US $5**

## Key Features Deep Dive

### Professional Watermarking

- Six preset positions (corners, centers, sides)
- 1-50% size scaling relative to image
- 10-100% opacity control
- Configurable margins
- Default: bottom-center at 14% opacity

### Smart Batch Renaming

- **Original**: Keep existing names
- **Sequential**: `Party_001`, `Party_002`, etc.
- **Event-based**: `Birthday_2025-08-07_001` with auto-dating
- Preview mode with conflict detection

### Technical Highlights

- **Proportional scaling** preserves aspect ratios
- **Background fill** for non-square images
- **macOS tag integration** for selective processing
- **Robust error handling** prevents batch crashes
- **Configuration persistence** saves your preferences

## Working with AI

The development experience was:  
\- **Iterative prompting** \- start with context → ask for a plan → zoom into code  
\- **Git co-authoring** \- each commit clearly co-signed for traceability  
\- **Security awareness** \- Claude flagged unvalidated inputs and proposed tested fixes  
\- **Cost efficiency** \- pay-as-you-go beats monthly subscriptions for small OSS bursts

## Impact & Usage

Super Fiesta now completely replaces my Photoshop macro workflow and Squash GUI for event photography. What used to take 30+ minutes of manual processing now happens in seconds with a single command:

```bash
python sf.py /path/to/photos --watermark logo.png --rename event

```

## Open Source & Future

The project is [fully open source](https://github.com/clawfire/super-fiesta/?ref=thibaultmilan.com) under MIT license. Planned features include:  
\- Tiny Tkinter GUI for non-CLI users  
\- Windows support expansion  
\- Multi-format output options  
\- Custom dimension presets

Perfect for event photographers, social media creators, and anyone who wants Instagram-ready photos without heavyweight tools.