Nanobonana Agent Public — AI skill for Claude Code

⚠️ Tento repozitář byl přesunut

**Aktuální verze žije na [github.com/faborsky/nanobanana-app](https://github.com/faborsky/nanobanana-app).**

Tenhle repozitář je archivovaný a **už nefunguje** — modely, na které se odkazuje (`gemini-3-pro-image-preview` a `gemini-2.0-flash`), Google v červnu 2026 vypnul.

Nová verze má aktuální modely, hlášení ceny za obrázek, dávkové zpracování, testy a **přibalený skill pro Claude Code**. Pokračuj tam: **→ https://github.com/faborsky/nanobanana-app**

---

Nanobanana

CLI tool for AI image generation powered by **Google Gemini 3 Pro Image** model. Designed to be orchestrated by AI coding agents like [Claude Code](https://docs.anthropic.com/en/docs/claude-code), but works great standalone too.

Supports single image generation, batch processing from JSON job files, visual references, and image-to-prompt analysis.

**Course material** for [Vibe Coding for Marketers](https://vibecodingformarketers.com) — a hands-on course teaching marketers to build real tools with AI.

Features

• Single generation — generate one image from a text prompt
• Batch processing — generate multiple images from a JSON job file with progress tracking
• Visual references — use a reference image to guide style and content
• Image analysis — describe an existing image to create reusable prompts
• Flexible output — choose aspect ratio (10 options), resolution (1K/2K/4K), and format (WebP/PNG/JPEG)
• Retry logic — automatic exponential backoff on API failures
• Graceful interrupts — Ctrl+C stops batch processing cleanly after the current job

Requirements

• Python 3.11+
• Google Gemini API key (free to get — see below)

Getting your API key

1. Go to Google AI Studio
2. Sign in with your Google account
3. Click "Create API key"
4. Select or create a Google Cloud project (the free tier is sufficient)
5. Copy the generated key — it starts with AIzaSy...

T