ck:ai-artist

To generate AI images from a concept, you run a local Python script that selects curated Nano Banana prompts via BM25 lookup and executes Gemini image models to produce styled outputs. You can optionally apply creative remix or wild transformations for variations.

The validation interview is a mandatory step collecting your preferred style, mood, colors, and aspect ratio to refine the final Nano Banana prompt. You can bypass this interview and proceed with default settings by providing the --skip flag.

Yes, you need a Python environment to execute the local scripts and the google-genai dependency to interface with Gemini models. The image generation workflow relies on these scripts to perform BM25 prompt search and call the Nano Banana models.

Yes, the generator uses 129 curated prompts that you can match to your concept in search mode, or remix and wild-transform them for creative variations. This allows flexible prompt-driven image creation across different modes for banners, characters, and stylized scenes.

The best way to create a tech conference banner is to run the generator script with your concept, specify the output file, and set the aspect ratio. The script uses BM25 prompt lookup and the Gemini model to generate the image with your configured dimensions.

Yes, image creation via Nano Banana supports configurable aspect ratio and quality mode. You can adjust these parameters when running the local Python scripts to ensure the generated output meets your specific visual requirements.