An AI request failed with "used its whole output budget reasoning and returned no text". What should I do?

The selected model is a reasoning model and it spent the entire output allowance of that request thinking, so nothing was left for the text. This can happen on short requests such as title suggestions or chapter summaries. Either pick a non-reasoning model for that book in Book details (under "AI engine for this book" and the model selection in Settings), or keep the model and retry; the message replaces the older, misleading "Empty response".