Which API key do I need for chapter illustrations? I write my books with Ollama.

From version 1.9.22, chapter pictures are drawn with an OpenAI or a Google Gemini API key (OpenAI is used when you have both). The chapters themselves can be written by any engine, Ollama included: your writing engine only picks the moment to draw, and the image key draws it. Add the key in Settings. Each picture is one image request on that key. Note that Google's free tier may not include pictures: if Gemini refuses with a quota message, turn on billing for the key in Google AI Studio, or add an OpenAI key.