#AI game with directus
37 messages · Page 1 of 1 (latest)
Interesting, but it seems that there is still a but when trying to start as when clicking on start, it quickly shows a different ui, but then goes back to the login screen.
i have the same 😓
Thanks for letting me know @real anchor and @open wharf . The issue you had should be all fixed now. Hope you'll retry!
works now 🙂
its really nice and i feel like it would be a perfect fit for story writing
this is because you can also tell it what should happen next
for example like this
i like the style tho 🙂
Haha, also tried imagining a dragon out of thin air but failed miserably. How are you doing this @south hill? Using OpenAI functions or something?
Would be a nice feature for alot of AI Bots, i did search for things like that for help desk bots but couldent find anything.
Would be nice if you could share some explaination 🙂
It’s actually much simpler than that… I have a generic “if this doesn’t make logical sense then…” in the system setup message. Chat GPT does a reasonable job at understanding that. You seem to have found a loophole though! 😅
Oooh. I never saw it justify why you can’t do something in my testing! That’s cool 😎
you cant do that much with it tho so its fine 🙂
too bad that there is a daily limit but i can understand that
It’s cool that you kinda made it part of the story and it worked but @real anchor randomly introduced something that wasn’t there and so it didn’t allow it.
And yeah, daily limit’s to protect my openAI a token balance! Had quite a few players over the last few days but will look at tweaking limits once I have more usage data.
Might allow byo api key at some point.
i fully understand 💪
If you play for a couple of days the story will eventually end. I’m working on some better mechanics around that.
imagine some image generation for every prompt so you have a visual (probs hard to do because the AI will always generate a differend environment)
cool design!
Exactly. And text to speech reading the text narrator style.
it would probably be hard unless you use some kind of prompts to see what should make the noice, so if its for example Lisa (some character you met) it will be her voice every time
Yeah. I was just going to do one voice as the narrator rather than each character but maybe it’s worth a try. I stopped after a couple of tests because the responses took so long. And the text to speech web Sdk sounds pretty average.
Ooooh! This could help. I’ll have a play. Thanks.
Only bad thing is that its not in JS 👀
Yea generation times are bad with some ai stuff
I like it how midyourney gives you the preview + status %
Sure Kevin! I’d be happy to! Let me know any details and I’ll write something up for you.
Would be awesome to just get an overall view of how it was put together, along with maybe a couple of code snippets if they are of note.
If you fill this in we can frame what the post might look like: https://github.com/directus-community/guest-authoring/issues/new?assignees=phazonoverload&labels=&projects=&template=new-post-proposal.yml
Thanks Kevin, submitted!
Great. I’ll respond on Monday. AFK today
David Shapiro's ACE_framework might help as well, It helps with layers, so can have a global goal for the app and then maybe more specifics for the charactors.
https://github.com/daveshap/ACE_Framework
having fun with it, giving myself magical powers and all.
Thanks. I’ll check this out.