CLI tool that generates images from text prompts using Google Gemini and renders them directly in your terminal using ANSI 256 colors and Unicode half block characters.
- Text to image generation via Google Gemini API
- 256 color terminal rendering with half block characters
- Interactive mode with animated logo and live settings
- Real time filter, texture, posterize and palette controls
- Custom palettes via PNG LUT files (create your own in Photoshop)
- Polaroid style border with optional caption
- Save original image, 256 color preview and JSON export
- macOS or Linux
- Python 3.9+
- A terminal with 256-color support
- A Google Gemini API key (get one here)
git clone git@github.com:hooop/axon.git
cd axon
pip install -e .Set your API key:
export GOOGLE_API_KEY=your_key_hereaxonLaunches the full interface: animated logo, configuration prompts (size, polaroid border), then a text prompt. After generation, navigate the settings menu with arrow keys to adjust the rendering in real time.
axon "a picnic on the grass, impressionist style, pixel art"Options:
--width N Terminal columns for rendering (default: auto-detect)
--size N Generated image resolution: 512, 768, 1024 (default: 768)
--pola Add a polaroid-style white border
--caption TXT Caption text on the polaroid border
After image generation, use arrow keys to tweak the rendering:
Filter: Silk (Lanczos) / Soft (Bilinear) / Crisp (Bicubic) / Raw (Nearest)
Texture: Clean (none) / Grain (Floyd-Steinberg) / Grid (Bayer ordered)
Poster: Off / Light (4 levels) / Heavy (2 levels)
Palette: None / custom LUTs from palettes/ folder
Press Enter to confirm, then choose to save and/or export.
- Open
axon_palette_256.pngin Photoshop (16x16 grid of all 256 ANSI colors) - Apply any adjustments (curves, hue/saturation, color balance...)
- Save as PNG in the
palettes/folder - The palette appears automatically in the menu next time you run Axon
