Feedback

I am often changing POV scene to scene. When I need to set it on a new scene, the process is: Actions->Custom POV Select Character Pick the POV Save I would prefer to have a dropdown like I do for labels where I can just set an existing POV directly. Maybe with a recent list of POVs I've set before since that's likely to be much smaller than the number of Character codex entries.

A search-and-REPLACE option in Codex so we can change the names of characters / objects / locations universally across multiple entries with one click. It would save a lot of time instead of manually updating each separate Codex entry.

One of my POVs has a slight plurality, so I set it to the default. That means any scene that doesn't show a POV is going to use my default. I would like to see that in the UI in a more obvious way. Some thing like: Default POV: 3rd Person (Limited) - John Smith

It would be great to have a section in each chapter for adding quick notes about the chapter. This is different from summarizing the chapter. A little section for adding any quick ideas or thoughts about the chapter. Useful for editing, revising and plot continuity, among other things.

Summary Please add support for message editing, response retry, and conversation branching within the chat interface. These features are standard in many modern LLM front-ends and greatly improve iteration speed, experimentation, and overall usability. When interacting with the LLM, users often want to: * Correct or refine a previous prompt without rewriting the entire conversation. * Retry the last model response to explore alternatives or fix generation errors. * Fork a conversation into a new branch to explore different directions without losing the original context. Currently, the interface does not provide edit/retry/branching tools, which forces users to copy/paste or restart flows manually. This disrupts workflow and makes iterative prompting much less efficient. Requested Features Edit Previous User Messages Allow the user to modify a past prompt. Automatically regenerate downstream messages based on the edited prompt. Provide clear visual indicators that a message has been edited (like “Edited” tags). Use cases: Fixing typos or unclear phrasing. Updating system assumptions or constraints without resetting the whole session. Retry / Regenerate Last Response Add a “Retry” or “Regenerate” button on model responses. Generate a new answer using the same user message. Use cases: Exploring multiple variations of a response. Recovering from hallucinations or errors without retyping prompts. Branching Conversations Enable users to “branch” from any previous message into a new conversation thread. Each branch should be treated as its own isolated session with inherited context up to the branching point. Provide a sidebar or tree view of branches for easy navigation. Use cases: Trying multiple story directions. Evaluating different solutions in technical problem-solving. Preserving original context while exploring alternatives. Benefits * Dramatically improves iterative prompt development. * Aligns with UX patterns established by major LLM front-ends. * Reduces friction for complex workflows like writing, coding, research, and game design. * Helps both beginners and power users experiment more fluidly. Conclusion Adding edit, retry, and branching features would make the chat interface more powerful, flexible, and user-friendly. These capabilities have become expected norms in leading LLM interfaces, and their addition would significantly enhance the user experience.

I can assign a category color to custom categories and set custom tags but not the default ones? Why? Let me do this. Right now "Locations" to all be red coded I have to make a custom "Locations" which defeats the purpose of the default.

In the font picker it would be great to have a small selection of handwriting fonts like Comic Neue, Caveat, Architect's Daughter, etc. I prefer that type of font when I'm drafting and switch to a sans serif when I edit.

There is currently no way to view your already existing thumbnail images in your codex, which is currently a very small thumbnail. It would be very useful to reference your thumbnails while writing by enlarging them when you click on them, just like you are able to reference your text.

Would love to see a split scene command. Very often I'll create a scene and fill it with beats, only to realize half way through writing the scene that it really should be two scenes. I'd love to be able to do a "/" command in writing mode that simply splits the current scene into two scenes. Everything above the command remains in the current scene, everything below the command is split into a new scene that contains all the remaining beats, prose, content, etc.

Can we have the option in the character sheet to include the different relations of the characters to each other? Like: Brother, Mother, Father, Cousin, Best Friend, Girlfriend,.... So the AI knows how the characters' relations are. And with that have a family tree function? Overview of characters and how they relate in a chart. If it makes little sense, please delete the feature request :D