generate-image

To generate scientific illustrations from text prompts, provide a natural-language description of your diagram or schematic. The script sends the prompt to OpenRouter image-capable chat completions and saves the returned base64-encoded image to a PNG file.

Yes, an OpenRouter API key is required to edit existing images. You must supply the input image along with targeted edit instructions, and the script processes the revision by routing the request through OpenRouter image-capable chat completions.

Yes, you can use FLUX or Gemini models for scientific diagram generation. The script supports model flexibility via OpenRouter, allowing you to select between Gemini and FLUX models to balance quality and speed for your visual outputs.

The best way to iterate on scientific figures for posters is to first generate the base visual from a detailed prompt, then apply targeted edit instructions to the saved PNG file for revisions like changing colors or adding and removing elements.

No, this image generation method only outputs PNG files. The script specifically decodes the base64-encoded image returned from the OpenRouter API and saves the final scientific illustration or edited visual directly to a PNG file.

A key limitation when creating infographic-style visuals with AI is the dependency on text prompt clarity for scientific accuracy. Additionally, all outputs are constrained to PNG format and require routing through OpenRouter, meaning API access is mandatory for generation.