Rendered at 23:56:19 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
MrRowTheBoat 12 hours ago [-]
Hi everyone!!! Thanks so much for the support.
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!
MrRowTheBoat 12 hours ago [-]
We are at $100 usage since posting this haha. I will probably cap it at $500 for you guys
dmd 12 hours ago [-]
It would be nice to be able to play with one's own account, so it's not billing you.
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...)
MrRowTheBoat 12 hours ago [-]
Yes!! This was a plan I had, BYOK, Bring your own key
I may do this today
cs1996 10 hours ago [-]
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.
MrRowTheBoat 10 hours ago [-]
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."
sphars 9 hours ago [-]
"Drink verification can to continue"
MrRowTheBoat 9 hours ago [-]
"Please slurp as loud as possible into the mic"
andrewfromx 8 hours ago [-]
[flagged]
ramesh31 11 hours ago [-]
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.
senordevnyc 6 hours ago [-]
Dude, just wanted to be sure you saw his comment about exposing your API key. You do not want to do that!
MrRowTheBoat 6 hours ago [-]
I think I replied to this, ephemeral secrets are minted per browser
panorama 13 hours ago [-]
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.
sorry :( I will think about this some more, the only reason I have to lock people out is because I can't afford the current costs lol
twosdai 6 hours ago [-]
Byok or monetize it. Seems like a great game. :)
MrRowTheBoat 4 hours ago [-]
DONE
smusamashah 12 hours ago [-]
Please add a demo video/voice where you show someone asking questions and AI answering it. I wanted to see what it is like but dont want to signup.
MrRowTheBoat 12 hours ago [-]
I will do this, I was not sure how to do this cleanly, my recording I did was a little awkward haha
MrRowTheBoat 4 hours ago [-]
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.
MrRowTheBoat 4 hours ago [-]
I also added a donation button, no donation is required, but I have had a few people request this to assist with the costs.
arexxbifs 15 hours ago [-]
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.
ChrisGreenHeur 14 hours ago [-]
So it was very life like and realistic.
arexxbifs 14 hours ago [-]
Not really, since it was something that particular character would clearly have known about and wouldn't have gained anything from lying about.
ChrisGreenHeur 14 hours ago [-]
Again, sounds very realistic.
MrRowTheBoat 12 hours ago [-]
This is solved by having a shared knowledge base between the agents!
JoeDaDude 9 hours ago [-]
I have an LLM running locally on my machine complete with STT and TTS. While I don't expect much, it will be fun to get my LLM to play the game.
MrRowTheBoat 9 hours ago [-]
hahaha, omg. I'd love to see a video of this
b800h 18 hours ago [-]
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.
MrRowTheBoat 12 hours ago [-]
Hiiiii, i would appreciate your feedback now that I resolved the openai issue. I think you will be surprised. I was!
wfvr 16 hours ago [-]
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.
wfvr 12 hours ago [-]
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".
MrRowTheBoat 9 hours ago [-]
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.
grahamburger 18 hours ago [-]
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!
bartleeanderson 18 hours ago [-]
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.
MrRowTheBoat 12 hours ago [-]
OpenAI credits ran out! Resolved!
b800h 18 hours ago [-]
Same.
camillomiller 17 hours ago [-]
Vibecoding at its finest
BenzeneDream 9 hours ago [-]
Useless comment of the day
lorenzo95 19 hours ago [-]
sounds neat, would love to try it but I don't make accounts for websites anymore.
MrRowTheBoat 12 hours ago [-]
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
garbagewoman 18 hours ago [-]
so if you got locked out of hn, you would just never comment again?
what a strange comment
yoz-y 17 hours ago [-]
I would assume it implies “websites that have no need to actually have an account”.
Although most time I cave in, I really understand the sentiment of the parent.
watwut 16 hours ago [-]
It makes perfect sense. You are being intentionally obtuse.
garbagewoman 15 hours ago [-]
setting an arbitrary point in time where you no longer will sign up for accounts does not, in fact, make perfect sense to me.
aljgz 19 hours ago [-]
[flagged]
BaselAshraf81 7 hours ago [-]
Consider using a model like supertonic 3 for tts.
Its light weight and can run on web inference.
ShinyLeftPad 16 hours ago [-]
Good way to get more people's voices into training dataset
ryanjshaw 14 hours ago [-]
Is getting people’s voices difficult? Doesn’t YT have more than enough data already?
MrRowTheBoat 12 hours ago [-]
I have no idea how to do that, but yes, as someone else noted, OpenAI might!
ShinyLeftPad 10 hours ago [-]
You can assess if model provider gets any identifiable data about the participating person though. Is that the case?
MrRowTheBoat 9 hours ago [-]
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.
ShinyLeftPad 9 hours ago [-]
I think it does thanks!
amelius 14 hours ago [-]
Maybe run a STT -> TTS filter in front of your mic?
MrRowTheBoat 9 hours ago [-]
I've built this pipeline before on my first iteration, because it's cheaper, it dooes add a lot of latency :(
ShinyLeftPad 14 hours ago [-]
I guess could work, real mic -> stt -> tts -> emit from a virtual audio device configured as the mic...
amelius 14 hours ago [-]
I only hope latency is not a problem.
wartywhoa23 15 hours ago [-]
The only relevant comment!
qwertox 14 hours ago [-]
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.
ShinyLeftPad 14 hours ago [-]
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.
13 hours ago [-]
ethannam9022 17 hours ago [-]
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.
garbagewoman 17 hours ago [-]
Maybe it only works for hn accounts created yesterday
MrRowTheBoat 12 hours ago [-]
Oops, ran out of OpenAI credits
MrRowTheBoat 8 hours ago [-]
okay, I posted my goofy prototype demo from a couple days ago.
donpdonp 18 hours ago [-]
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.
londons_explore 18 hours ago [-]
It was working great... But now it says it can't connect.
Did the server go down or did I hit one of your limits?
MrRowTheBoat 12 hours ago [-]
Ran out of OpenAI credits, resolved now
dbx842 6 hours ago [-]
This is really fun, great work!
MrRowTheBoat 4 hours ago [-]
Thanks <3
anonzzzies 14 hours ago [-]
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.
aswegs8 7 hours ago [-]
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.
MrRowTheBoat 3 hours ago [-]
I’d love to hear your improvements
fitsumbelay 7 hours ago [-]
Brilliant
october8140 12 hours ago [-]
AI: Yea I did it.
cindyllm 12 hours ago [-]
[dead]
garbagewoman 18 hours ago [-]
"voice connection failed"
zerratar 17 hours ago [-]
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.
kristophph 17 hours ago [-]
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.
MrRowTheBoat 9 hours ago [-]
[dead]
MrRowTheBoat 12 hours ago [-]
Sorry! Ran out of OpenAI credits, resolved now
qmmmur 12 hours ago [-]
Boo. AI Slop. Be critical of what you’re making please.
thrance 14 hours ago [-]
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.
ZeroDayDreamer 10 hours ago [-]
[dead]
rajivnaskar 11 hours ago [-]
[dead]
wundervault 5 hours ago [-]
[dead]
ksingla025 16 hours ago [-]
[flagged]
mohammedmsgm 15 hours ago [-]
[flagged]
runtime_lens 17 hours ago [-]
[flagged]
ccforai58 16 hours ago [-]
[flagged]
wewewedxfgdf 18 hours ago [-]
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?
9dev 16 hours ago [-]
The entire game hinges on the premise of being able to narrate and understand dynamic text on demand. Pre-recorded messages don’t work for this.
MrRowTheBoat 9 hours ago [-]
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
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.
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".
what a strange comment
Although most time I cave in, I really understand the sentiment of the parent.
Its light weight and can run on web inference.
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.
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"