I launched this at midnight prior to logging off and going to bed, funnily enough a massive storm hit us around 3am and I woke up and noticed my OpenAI account failed to fund lol. So we ran out of funds and it caused things to break for some of you.
Just waking up here on the East coast, once my daughter is occupied I will try to answer more of your comments!
Instead of that, I would not put in my own key, you should get a brand partnership. Something like the Liquid Death Murder Your Thirst water brand. https://liquiddeath.com This would be perfect for them and paying for all the AI use would be nothing compared to their normal advertising. Win win.
Are you proposing a small add on the site? or does the AI agent do something like this:
"Hey give me a second to answer while I take a sip of this crispy liquid death, these killer cans of soda-style sparkling will attack your tastebuds with soda-level flavor and only 1/10th of the sugar."
Maybe play around with CPU inference on smaller models too. Frontier models are probably overkill for simple language responses at this point; I suspect you could get pretty good results with Qwen doing this, and Kokoro TTS is really good now.
Very neat idea, and overall it is more well-done than I assumed a vibe-coded game would be. Unfortunately I ran out of time and now the game locks me out of doing anything and I don't know who the murderer is :( I wish there was just a way to have it revealed.
I've always thought that a game like Civilization should allow the user to interact with different opposing world leaders, but in a non-deterministic way as opposed to a subset of fixed interactions. Given what you know about this leader's temperament, goals, and so on, you can conduct diplomacy in a way that can be way more immersive and crafty than what the game currently gives you, which hasn't seem to evolve much in decades.
Diplomacy is great but computationally suffers from deterministic outcomes - some starts are better than others, there are actual optimal moves to make, you can (and should) play to a stalemate, etc. which take away from the fun - playing against the best moves to make vs against a crafty person.
There is a new Diplomacy card game (The Golden Blade) that is that is shorter and more approachable (plus no player elimination) and not solvable. You still play secret orders; it is just that now your "borders" are your left and right opponent. You can still work together for extra benefit or betray your neighbor and backstab them. Obviously, there is a lot less potential for complex tactics each turn but it still gives some of the vibes of the classic in a format that is easy to get to the table. There are a lot of YT videos from content creators that you can get a sense of how it plays.
Source: I designed it :) (sorry for the minor self promotion)
I played various murder mystery games this summer and one of them had this feature: you could call a suspect and ask arbitrary questions for a limited amount of time (three minutes or something). The problem was of course that the model hallucinated an answer to one of the questions, suggesting that other suspects were related by blood even though they clearly weren't.
Even the tag-line, "one poisoned patriarch, one locked-in truth" sounds very AI, and makes me think that I'd struggle to enjoy this. But I'll give it a go, because the idea is excellent.
I think this would work very very well in combination with human writers.
Congrats, I have not had the chance to try it yet, but what I saw looked awesome and I'll give it a go, for sure. I'll update this comment when I do.
I've been working on a similar concept for 6 months now (https://game.selna.ai, portuguese only for now, sorry!), and it's been a journey. I've been working with VR professionally for some 8 years now and creating what I call a "social virtual reality" seemed simple enough, but has turned out to be much more complex than I initially imagined.
In the same way that in normal VR a single badly scaled object can break immersion, in "social VR" a single contradiction kills the experience. But when stars align and eveything works, it's magical.
I started with whodunnits but have expanded it into linear and non-linear storytelling. At this point the generation pipeline has some 40+ separate steps, the output of ones feeding others, until it all comes down to single prompt the NPC uses to talk using the realtime API.
The platform can create a complete narrative (including cast personalities and images) in some 20 minutes, and in the process generates the equivalent to a big book of text.
I've been working on this just because I can and I am a bit obsessed with getting it working perfectly, and I don't really know what I'll do with it when it does.
Sorry for the self-reply, I couldn't figure out how to edit the comment. I tried it now, it's very good. I love the fact that you can just move from one suspect to the other quickly.
You'll need to generate and give the characters a canonical current date and time, that will fix a lot of different subtle issues.
You have a huge break of character or the fourth wall whenever the player makes an accusation, the character goes like "you're making a serious claim, let me formally validate it before I respond" and then they refuse to reply to anything else while "the response is being validated".
Love it, and I think this is the future for RPG gaming. Endless immersion. I see this project expanding into "experiences" or "scenarios" as well, but I like to focus on small things first. If people like this, I will build more. If it's a cool demo, but not worth putting more time into, I'll probably just move on.
Neat! I've been wanting to make something like this for a long time, well before LLMs came around. It was one of the first things I tried out with an LLM in the early days of ChatGPT. My son and I made little mysteries for each other and just pasted them in a chatgpt conversation with instructions for it to pretend to be the other characters and whatnot, then handed the keyboard to the player. Mine didn't work that well, I don't think he ever solved it; partially because chatgpt kinda led him astray. When I started his, I just point blank asked the mystery and it told me the solution. I felt bad about that, he had put a good amount of effort into it, but I had reminded him to tell it not to reveal the answer and he didn't. Anyway, I figured someone would make this eventually! Excited to give it a try!
In an effort to keep things running longer, before hitting my budget of $500, I am reducing times to 15 minutes, and I've added BYOK.
I don't think I can make everyone happy with the solution, your key is saved on your own browsers localStorage, I do not save or store it in my database. I suggest everyone creates a new key dedicated to our service so you can invalidate it later on. That key is sent to our api using `x-openai-key` header and then we mint ephemeral client secrets to perform web rtc functions.
Really like this idea and the way you implemented it is very interesting to read about. BYOK is definitely needed as people will want to use this with their own account. Genuinely novel and nicely done. Keep it up and thank you!
This is a really fun take on a detective game. One small thing I noticed: during the initial detective briefing, the time always seems to show 0:00. Not sure if that’s intentional or a bug.
sounds good, but it won't get past Dispatch confirmation. You should specify what networking permissions or whatever is needed. My It goes red and the text to read disappears.
Very understandable… I originally didn’t have accounts when sharing it to friends. The account is really just there as a barrier to try and circumvent bots and bad actors as much as possible, even though there are still flaws. The goal was simply to have a way to force a max timer on someone so that more people can try this out
The authentication service, Clerk, will gain your First name, and Email address if you use Google SSO.
The model provider does not have any identifiable data such as emails, user ids, first name, nothing at all related to your User account. it's the agent context, and the user who is identified in prompt as "DETECTIVE". I don't know if this answers your question.
Not really. It is true in a way, but the question is if this game is made to collect data, or just to present a game, where the recordings are then deleted, in good faith.
The developer's homepage gives the impression that this is more for the tech and fun aspect, rather than collecting speech samples.
this uses a proprietary gpt model and so the data is sent to closedai. even if the developer is in good faith you can expect your voice to be part of training data.
even without trying it, its so cool this exists. I think this kind of voice interaction has been a dream of adventure game creators for so long, and now its possible. there was a game many years ago (in the days of messengers like trillium being popular), I forget the name, where one of the characters would make a landline phone call to you. The voice acting was great and it felt novel and exciting to see it work.
It is a nice idea ; I will try it later. Everyone can poop on AI, that’s fine, but this specifically states it is AI and how it is used, so just click on another HN link to cry about AI, not the ones that actually specifically mention it.
Wow!!! That's awesome. So much AI slop. That here is definitely not. Kept me hooked, even though I am not a crime nut. Much improvements possible for sure, but I can see how this becomes a serious game with a lot of players.
Meh, it's like the fourth such "interview AI to solve the mystery!" implementation I've seen. It's always so bland, uninspired and boring, and drags on for waaay too long.
same here. js error though is 0oxvw2ytjke1_.js:1 POST https://api.openai.com/v1/realtime/calls 429 (Too Many Requests). and auth bearer token in clear js request here towards openai. someone could have just used the token for their own good. cant go directly to openai's apis from the website. huge security risk.
True, I have a dungeons and dragons combat tracker (https://topoftheround.com) where I let the user generate loot items with different settings for them like rarity or magical/non-magical etc. and I am using a firebase cloud function that talks to the actual AI and returns the result. I also have set a limit to how much that token is allowed to spent per month, hopefully OP has thought of that as well. Interpreting and generating voice probably uses a lot of tokens.
I have no problem with AI generated text and story but I just can't stand AI voices. Even a hint it is AI narrated and I check out. I would encourage you to consider getting it voice narrated by real people. I'd rather listen to bad/amateur human narration than AI voices.
As a second point - this is very interesting using voice control as game input are there viable alternatives for realtime voice uasing your own GPU?
I think t's possible to use static voice, but it would create a very rigid experience, you would have to have a ton of generic sounds prepared and identify when to play each one. The fun of this is the fluid conversation. I can imagine the static responses being pretty terrible for unrecognized questions. LLMs provide really good RP.
I launched this at midnight prior to logging off and going to bed, funnily enough a massive storm hit us around 3am and I woke up and noticed my OpenAI account failed to fund lol. So we ran out of funds and it caused things to break for some of you.
Just waking up here on the East coast, once my daughter is occupied I will try to answer more of your comments!
I know there's a way to do this with A\ powered apps like this, I'm sure there must be for OAI?
(By the way, you're exposing the key. So anyone can just use up that key for their own nefarious purposes...)
I may do this today
"Hey give me a second to answer while I take a sip of this crispy liquid death, these killer cans of soda-style sparkling will attack your tastebuds with soda-level flavor and only 1/10th of the sugar."
I've always thought that a game like Civilization should allow the user to interact with different opposing world leaders, but in a non-deterministic way as opposed to a subset of fixed interactions. Given what you know about this leader's temperament, goals, and so on, you can conduct diplomacy in a way that can be way more immersive and crafty than what the game currently gives you, which hasn't seem to evolve much in decades.
https://boardgamegeek.com/boardgame/483/diplomacy
Source: I designed it :) (sorry for the minor self promotion)
I think this would work very very well in combination with human writers.
I've been working on a similar concept for 6 months now (https://game.selna.ai, portuguese only for now, sorry!), and it's been a journey. I've been working with VR professionally for some 8 years now and creating what I call a "social virtual reality" seemed simple enough, but has turned out to be much more complex than I initially imagined.
In the same way that in normal VR a single badly scaled object can break immersion, in "social VR" a single contradiction kills the experience. But when stars align and eveything works, it's magical.
I started with whodunnits but have expanded it into linear and non-linear storytelling. At this point the generation pipeline has some 40+ separate steps, the output of ones feeding others, until it all comes down to single prompt the NPC uses to talk using the realtime API. The platform can create a complete narrative (including cast personalities and images) in some 20 minutes, and in the process generates the equivalent to a big book of text.
I've been working on this just because I can and I am a bit obsessed with getting it working perfectly, and I don't really know what I'll do with it when it does.
You'll need to generate and give the characters a canonical current date and time, that will fix a lot of different subtle issues.
You have a huge break of character or the fourth wall whenever the player makes an accusation, the character goes like "you're making a serious claim, let me formally validate it before I respond" and then they refuse to reply to anything else while "the response is being validated".
I don't think I can make everyone happy with the solution, your key is saved on your own browsers localStorage, I do not save or store it in my database. I suggest everyone creates a new key dedicated to our service so you can invalidate it later on. That key is sent to our api using `x-openai-key` header and then we mint ephemeral client secrets to perform web rtc functions.
what a strange comment
Although most time I cave in, I really understand the sentiment of the parent.
The model provider does not have any identifiable data such as emails, user ids, first name, nothing at all related to your User account. it's the agent context, and the user who is identified in prompt as "DETECTIVE". I don't know if this answers your question.
The developer's homepage gives the impression that this is more for the tech and fun aspect, rather than collecting speech samples.
Its light weight and can run on web inference.
Did the server go down or did I hit one of your limits?
As a second point - this is very interesting using voice control as game input are there viable alternatives for realtime voice uasing your own GPU?
"Tell me where the knife was hidden?"
"I don't understand your question"