nano-banana-photoshoot

KO | EN

gitlite — search

#javascript #typescript #go #android #machine-learning #rust #docker #java #awesome #cli #awesome-list #pytorch #security #openai

nano-banana-photoshoot

(★ 2)

Nano Banana 2 Photoshoot Prompt Engine -- A Claude Code skill that rewrites plain-English image descriptions into structured JSON prompts with cinematic lighting, lens settings, composition, and color grading, then fires them to Nano Banana 2 via API. Batch-process a 30-day social calendar from a spreadsheet. Every output looks professionally shot.

nano-banana-photoshoot 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Nano Banana 2 Photoshoot Prompt Engine

A Claude Code skill that turns plain-English image descriptions into structured, photo-shoot-quality prompts and fires them to Google's Nano Banana 2 (Gemini 3.1 Flash Image) API. Every output looks like it came from a professional photographer, not an AI generator.

## What It Does

You write: *"A detective in a rainy alley at night"*

The engine rewrites it as a structured JSON prompt with Rembrandt lighting, 35mm f/1.4 lens, shallow depth of field, chiaroscuro shadows, film noir color grading, and negative prompts to suppress AI artifacts. Then it calls the API and saves the image.

## Features

- **Prompt Engineering Pipeline** -- Rewrites plain text into structured JSON with lighting, camera settings, composition, color grading, style direction, and negative prompts
- **Batch Processing** -- Upload a CSV/XLSX spreadsheet of 30 days of social media posts and generate all images in one run
- **Platform-Aware** -- Automatically maps Instagram Feed to 1:1, Stories/Reels to 9:16, LinkedIn to 16:9, Pinterest to 2:3, and more
- **Dual API Support** -- Works with the native Google Gemini API or fal.ai REST API
- **Organized Output** -- Saves prompt JSON files, images, and a manifest CSV linking everything together
- **Dry Run Mode** -- Review all structured prompts before spending API credits
- **Style Presets** -- Built-in profiles for corporate, law enforcement, editorial, product hero, lifestyle, and cinematic looks
- **Photographic Reference Library** -- Lighting setups, lens selection guides, composition templates, and color grading presets

## Quick Start

```bash
# Set your API key
export GEMINI_API_KEY="your-key-here"

# Single image
python scripts/generate_one.py "Product hero shot of wireless earbuds on dark surface" \
  --platform instagram --api gemini

# Batch from spreadsheet
python scripts/batch_generate.py \
  --input social_calendar.csv \
  --output-dir ./output \
  --api gemini \
  --resolution 2K

# Preview prompts without calling the API
python scripts/batch_generate.py \
  --input social_calendar.csv \
  --output-dir ./output \
  --api gemini \
  --dry-run
```

## Project Structure

```
nano-banana-photoshoot/
├── SKILL.md                              # Claude Code skill definition
├── scripts/
│   ├── batch_generate.py                 # Batch spreadsheet processor
│   └── generate_one.py                   # Single image generator
├── references/
│   └── prompt-engineering-reference.md   # Photographic style deep reference
└── sample_social_calendar.csv            # Example spreadsheet format
```

## Output Structure

```
output/
├── prompts/
│   ├── 2026-03-24_instagram-feed_forensic-investigator.json
│   └── 2026-03-25_linkedin_command-center.json
├── images/
│   ├── 2026-03-24_instagram-feed_forensic-investigator.png
│   └── 2026-03-25_linkedin_command-center.png
└── manifest.csv
```

## Spreadsheet Format

Your CSV/XLSX needs at minimum an image description column. The script auto-detects columns named `image_description`, `description`, `prompt`, `visual`, or `creative`. Optional columns for richer prompts:

| Column | Purpose |
|---|---|
| `date` | Publish date (used in filenames) |
| `platform` | Target platform (sets aspect ratio) |
| `topic` | Theme context for prompt enrichment |
| `audience` | Target audience for tone calibration |
| `caption` | Post copy (provides additional context) |

## API Options

| Backend | Env Variable | Model | Notes |
|---|---|---|---|
| Google Gemini | `GEMINI_API_KEY` | `gemini-3.1-flash-image-preview` | Free tier available at [AI Studio](https://aistudio.google.com/apikey) |
| fal.ai | `FAL_KEY` | `fal-ai/nano-banana-2` | Pay-per-image, supports resolution selection and thinking levels |

## Requirements

- Python 3.10+
- `requests` (auto-installed)
- `openpyxl` (auto-installed if processing .xlsx files)

## Built For

Claude Code skill system. Drop the entire folder into your skills directory and Claude will automatically use it when you ask for image generation, social media visuals, or batch image processing.

## License

Internal tool. Not for redistribution.

nano-banana-photoshoot

nano-banana-photoshoot 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

nano-banana-photoshoot

nano-banana-photoshoot 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README