Feedback feature improvement

I’ve been using the Feedback feature extensively on full novel chapters, and there are a few changes that would make it significantly more useful to me as a long-form fiction writer:

1. Allow feedback on selected text only.
I would love the option to highlight a specific passage or scene and run Feedback only on that selection. Running Feedback on an entire chapter can be expensive in credits, particularly when I only want another look at a few paragraphs or one scene. It would also make the tool much more useful during revision, when I’m addressing specific problem areas rather than evaluating the chapter as a whole.

2. Show the credit cost before running Feedback.
Please show an estimated credit cost before the user submits the request. When working with longer chapters, it’s difficult to know how expensive a Feedback run will be until after the credits are spent. A simple “This will use approximately X credits” would make budgeting credits much easier.

3. Believability feedback needs more nuance.
This category can be very useful, but it sometimes identifies something as unbelievable without adequately considering the established circumstances, worldbuilding, character abilities, or information already provided in the manuscript. I’d like it to distinguish more carefully between an actual plausibility problem and something unusual that has already been supported by the story.

4. Feedback needs stronger adherence to established canon and breadcrumbs.
This is especially important in long-form fiction and series work. Feedback sometimes recommends explaining, changing, or questioning something that has already been deliberately established, foreshadowed, or answered earlier. Ideally, it should recognize existing canon and breadcrumbs before identifying something as a gap. Otherwise, following the recommendation can lead to unnecessary repetition and over-explaining things the reader already knows.

5. The cliché detector needs more nuance around physical and sexual responses.
In particular, please stop flagging normal physiological responses during orgasm or intimacy simply because similar language appears elsewhere in fiction. Humans have one nervous system with a finite number of ways it responds to stimulation. Trembling, tightening muscles, changes in breathing, loss of coherent thought, etc. are not automatically clichés. Sometimes they’re simply human physiology.

Or, put less formally: Robot, we have one nervous system. It responds in only so many ways. 😂

I’d love for the cliché analysis to distinguish between genuinely tired figurative language and unavoidable descriptions of normal human physical responses.

Overall, I find Feedback valuable, which is why I use it enough to notice these issues. Giving writers more control over what gets analyzed, greater transparency about credit use, and better awareness of manuscript context would make it much more powerful for novel-length work.