Probably just me not understanding how to use Narrova, but

Support
1

Really sorry if this is not the spot for this, I’ll move it if needed! I’ve loved working with Narrova, it has helped me elevate my story, digging out what was already there, fine-tuning scenes that were just shy of the mark, etc. As I get further into story development with it, though, I find myself reluctant to switch to a new conversation. I understand that context and narrative efficiency percentage goes down, but when I do start a new chat (and upload the previous conversation for context), the first few responses usually involve fixing or clarifying inaccuracies. I suspect this happens because the chat I upload includes errors—like when Narrova suggests something that doesn’t fit the story and I respond to refine it. To fix this, I’d just remove the incorrect suggestions, right? Easy enough. But it gets muddled when I’ve gone through multiple editing rounds on a pivotal scene that’s really threading the needle. I’ve tried editing the markdown file to keep only the relevant parts, but I keep coming back to: This is too much work, surely there’s a better way to do this. The context would be very helpful to keep continuity throughout new chats, but I think what I’ve ended up doing is filling it with some ‘slightly off’ responses that may have SOME good information in there that I want to keep. Sorry if this isn’t clear. I feel like I’m missing something or using this tool incorrectly. What would you suggest? Should I create a file with the key details—like lore, plot elements, and POV constraints—for story context? But when it comes to storyform context and continuity, my main issue happens when I start a new chat, upload a previous chat(s) for reference, and Narrova pulls random pieces of context. Sometimes it’s relevant, but other times it isn’t helpful. In one instance, it just used the title of a chat: about “reckless power sparking worry,” leading it to focus on my main character being reckless and worrisome in a vacuum, instead of pulling details from that specific conversation.

It also accesses older files that I’ve deleted. It still pulls ideas from those versions, and I assume I just have to wait for the system to refresh on its own. That’s likely on me for not uploading it straight into the chat, as I only recently realized how much more sense it makes for a real quick check on a draft scene (so that’s the level of Narrova Literacy I’m at).

In short, I’m hesitant to start new chats because the process feels messy, but avoiding it only makes things worse. What are the key steps or shortcuts I should be using to make this easier that I am definitely missing?

2

This is really great feedback - thanks so much!

The easy one first, if you’re deleting a document from context and yet its still being referred to, that would be on us, and I’ll look into it (let me know if thats what you meant).

For the bigger request, Muse (the previous version of Narrova) used to have this thing where you could “mute” responses individually–which I suspect would help out in your use case where its kind of gone off the rails, or said something not quite right – you don’t want it included in context, but you don’t necessarily want to start over with a new conversation.

This would be the equivalent of doing what you’re doing now, which is download, and then edit on your own, and then re-upload into context.

I haven’t brought back the Mute feature because technically Narrova isn’t built the same, so it would be more difficult to pull off, and more importantly – it’s much better and more precise doing what you’re doing now. I hear that it’s a pain for you, and it could be that there is a better way, but managing the context yourself for now will get you the best results.

There is another feature I have yet to add, which will automatically compact your conversation once it gets down to a low number in efficiency, effectively boosting it back up again. It may be that this solves your issue with context and new conversations. I imagine I’ll be able to get that out sooner than later (hearing this motivates me to do it sooner).

I’ve always been of the opinion that a new conversation/context reset is always a good thing – especially given the fact that things can get confused within the same conversation. It’s like building this platform, whenever I start a new feature I start a new coding conversation–which is just like starting a new conversation in Narrova - when I would want to work on a different part of the story–a scene, or a Storypoint, or a handful of Storybeats, I would just type /new and start on that.

One of the greatest value-adds for Dramatica is the concept of the Storyform. You can illustrate each and every aspect of the Storyform (450+!) in isolation, and yet the Storyform will tie it all together for you. It really is a grand context for narrative (which is why the open-source version is called Narrative Context Protocol!).

I’ll still add that compaction service in there, just know that keeping everything in context will lead to the same kind of result in does in humans - where you’ll end up saying the same thing over and over again, just with a slightly different context–which is not taking advantage of Dramatica theory and the expansiveness that exists when you can break out and explore all kinds of different avenues that end up being tied together in one grand argument.

3

One of the best easy solutions to the problem of unsolicited or not quite fit ideas getting in the way even after settling for a story idea and discarding other ideas.

And this too. I usually ask for summaries when I feel our conversation has been long enough to have some “noise issues.” We (users) could benefit from having such a summary and then editing it (some noise will always remain) as the reference document even if we delete other documents or the conversation itself.

Thanks for your constant effort. Narrova might not be perfect, but it’s still the best thing out there to help us craft better stories.

1 Like