Inworld GDC Ambassador Demo
1:23
4 ай бұрын
Indie Games powered by Inworld AI
2:42
New Feature Demo: Long Term Memory
3:24
Webinar on New Product Features
45:50
Пікірлер
@masturahmustafa3144
@masturahmustafa3144 Күн бұрын
Hi, my microphone is working on Unity windows but when i talk to the character it doesnt react to my voice.
@ZukunftBilden
@ZukunftBilden 3 күн бұрын
German is really missing.
@Hshdjdjdhd
@Hshdjdjdhd 3 күн бұрын
Stimmt :/
@user-nj3ej4ei7t
@user-nj3ej4ei7t 3 күн бұрын
Why would guys pick a song like that? What were you thinking?
@kayleewoz
@kayleewoz 5 күн бұрын
Fantastic news! I've been waiting for this for a looong time. Can Inworld characters be integrated into VRChat and Second Life yet? I have developed my own ChatGPT characters with TTS voice support in Second Life, but would live to use Inworld AI characters too 😊
@SignorCipolla
@SignorCipolla 9 күн бұрын
thanks I have a question..This game is safe for children or is like that obscene Second Life game? I hope this one dont have ANY sexy or nudity contents. Thanks for the info!
@Iamnotheretocompete
@Iamnotheretocompete 16 күн бұрын
can we make two characters speak with each other? in unity? can me use custom models of characters that we design in blender? can the characters move and adapt to the environments? like walk run fight etc
@gamesking8654
@gamesking8654 16 күн бұрын
Is it possible to integrate real system time and date for the npc? if so how do I achieve it, I have been trying to achieve it for my project. Thank you
@Iamnotheretocompete
@Iamnotheretocompete 17 күн бұрын
AI NPC engine no longer available for download
@moonytra4551
@moonytra4551 25 күн бұрын
The mod is no longer available on Nexus. Is there another way to access it?
@VarunSachdev
@VarunSachdev 29 күн бұрын
Seems like the option to add the Wikipedia link has been taken down. Can clearly see why it would be objectionable, almost on the lines of copyright probably. However, it would be such a cool option to continue to have.
@VarunSachdev
@VarunSachdev 29 күн бұрын
or is it to discourage developers from copy pasting? 🤔
@VarunSachdev
@VarunSachdev Ай бұрын
So exciting and enabling. Thank you Guys! More speed and Effectiveness to You :)
@zakaria_9m
@zakaria_9m Ай бұрын
HOW TO INSTALL IT 🤔
@WilliamVideoUT
@WilliamVideoUT Ай бұрын
Are you going to show how to do this in UE 5.4. as an extension of your " getting started" series episodes 1-4. I've entered my own goals and brain from my Inworld character to the "user" metahuman in the Inworld's Demo 2. But at my level of reliance on blueprint scripting in UE, the demo presents some difficulties for me. For one, I can't change Demo to a third person shooter. I have movement animations trigged by a 'trigger box game logic" that I'd like to activate with a player pawn,. I have managed to get Hair Grooms to remain on the metahuman, to put in my own environment and a few other things. But that is really just repurposing the Demo 2 game logic. In particular, I use the emotion arrays and ABP on the demo. Also, the metahuman is not exactly the same. I found the while I can import an metahuman from Quixel Bridge, it won't function "as is" in the Demo. Of course I may just be displaying my ignorance here, but that is why I would benefit greatly from a tutorial on goals and the attendant APB for body animation and the emotion arrays for the facial morphs. I will welcome an extension of your series that is presented at the same audience level or whom your first four episodes are focused. Is that likely to come anytime soon, or at all? I did try to ask this question on Discord a few months ago, but here it seems in the context of your tutorial for Unity, the question is posed much more clearly. I hope so.
@-TuberStriker-
@-TuberStriker- Ай бұрын
The thumbnail...
@VoidKaiser
@VoidKaiser Ай бұрын
like for the pentagram face behavior
@captquazar
@captquazar Ай бұрын
Actually not interested in Games, rather personal assistant avatars with good empathy and body language, but I like the 'customize your character' tech shown here. Localized AI processing here now (and rendering) with Nvidia RTX 4060+ series and above possible if you have an otherwise rather beefy PC. See "ChatRTX" now in alpha testing. Local processing is desirable if you don't want 'cloud overlords' looking over your shoulder. [note local AI requires a lot of CPU cores and Vram on the graphics card, at least 8gb minimum video ram.] Psych studies show 'normal' dialog is 5 seconds or less, so your local processes must be well integrated and efficient.
@Tyke18
@Tyke18 Ай бұрын
From 6:28 onwards when talking about the access token I think this is out of date now because the Studio access token method of adding a character is different in SDK 3.4.
@IrtazaUmer
@IrtazaUmer Ай бұрын
Hi, can you please guide me on how to make inworld work with the meta sdk for VR
@nagashud
@nagashud Ай бұрын
Is it me or some things changed about narrated action recently ?
@Norbyyyyy18
@Norbyyyyy18 Ай бұрын
I don't see all characters and scenes (only a few) when i login in Unreal with Inworld Studio. What to do?
@inworldai
@inworldai Ай бұрын
Do you have the latest update of the Inworld AI plugin? Are the given credentials the ones aligning with the workspace? If you continue to have any problems, please reach out to us at [email protected] for closer attention.
@cpypcy
@cpypcy 2 ай бұрын
Nobody would use technology in game that is limited to proprietary hardware. Especially if it's linked to such important thing as an AI. If it can run on any hardware then sure.
@TransmentalMe
@TransmentalMe 2 ай бұрын
Yeah, nobody would ever use DLSS or Nvidia Reflex. Nobody would think of using something as insane as dual sense haptic feedback for trigger pressure. It would be unheard of for someone to do something as crazy as make a game that relies on proprietary pressure sensitive buttons to determine if an action does something different... Buttons that were completely exclusive to a single console generation from a single brand. People have, do, and will. Whether it flops or not is the real question, because sometimes it makes a generation, other times it gets ignored like yesterday's news and destroys a company.
@inworldai
@inworldai 2 ай бұрын
To clarify: The point of the demo is to show the future of on-device AI. It's not to suggest people develop today for one type of hardware. It's expected that model size will decrease at the same time that the next generation of devices are created with greater processing power. At some point, those two trends will converge and on-device will be broadly practicable.
@alexinaboxx
@alexinaboxx 2 ай бұрын
@@inworldai​I think this is correct. It has all to do with fail states from cloud connection. I would love to develop with fully dynamic characters but many devs doing non-mmo may not want to count on the cloud for player progression if it breaks in real time. This needs to be on device to really catch on with the development community I think. Otherwise it would make more sense to voice and motion capture the performances traditionally. How much vram or TOPs would be needed to move all this to decide?
@DavideSpagocci
@DavideSpagocci 2 ай бұрын
would be awesome if ai can translate all videogames languages storytelling,actually this is a problem some games don't have language.
@inworldai
@inworldai 2 ай бұрын
We will be launching a multilingual feature soon and we are also making demos where language isn't so important. Look at our channel page for the AI Coop video Brian Cox posted. We also just demoed a similar experience with dogs internally.
@kira7x2
@kira7x2 2 ай бұрын
Looks like shit tbqh
@razz9266
@razz9266 2 ай бұрын
Skynet enter the chat The chat :
@lolmao500
@lolmao500 2 ай бұрын
Lets say tomorrow you would want this in real time locally on your CPU AI chip... how many TOPS do you need?
@ArthurTully
@ArthurTully 2 ай бұрын
it must be in the cloud. how they gonna charge you properly and make sure you didnt pirate
@inworldai
@inworldai 2 ай бұрын
The types of LLMs we use for our character conversations and orchestration are currently too large to be used on-device as on-device LLMs are still currently early days and experimental. This demo is showcasing the technical achievement of moving two models NVIDIA RIVA ASR and NVIDIA Audio2Face on-device to reduce latency and costs. LLMs, however, are continuing to get more powerful at smaller sizes and devices are continuing to be developed with more processing capabilities. It's expected that fully on-device AI will be possible once those two trends converge. That's the day we're working towards -- but the tech and ML techniques still need some time to get there.
@Rauzwel
@Rauzwel 2 ай бұрын
You really need to color the responses, I don't want to feel like I'm talking to a PR spokesperson. Allow swearing, rudeness, political incorrectness, bigotry, etc., or these will never feel like real people. There was slight rudeness at the beginning, but I'm not sold that current language models with their "safety measures" will ever be truly convincing. Imagine talking to a random bandit NPC in the Witcher 3, how do you think he'd sound? Not pleasant, I'm sure.
@inworldai
@inworldai 2 ай бұрын
This character is an executive the player is trying to pump for information in a detective game and so he's speaking in character. We don't allow hate speech of any kind but Inworld has configurable safety that allows you to turn off all safety guardrails around swearing, discussions of alcohol/drugs, violence (for gaming use cases where that's part of the world), and more. We want your Witcher 3 bandit to sound believable.
@Shieftain
@Shieftain 2 ай бұрын
@@inworldai Way to stifle creativity there with your over the top censorship. How are you supposed to create different kinds of villains if you're going to censor their motivations? I also noticed you deleted a comment from another individual that made a fair critique of your product. What are you so afraid of?
@inworldai
@inworldai 2 ай бұрын
Our safety systems allow for a number of in-game and in-character contexts where it makes sense for characters to discuss topics or have motivations other AI safety systems would flag as unsafe. However, we do not support the use of hateful slurs or characters created to incite hate. We do work with individual customers to enable use cases outside what we allow on a case by case basis given that they meet our terms of service and are not characters created to incite hate.
@Shieftain
@Shieftain 2 ай бұрын
@@inworldai Do you understand how incredibly subjective the term "hate" is? You're just asking for trouble when you put on restrictions like this. The first amendment in the US and similar laws about free speech in other countries protect you anyway, so placing such an arbitrary restriction is only to YOUR detriment. And you still haven't answered my question about why you deleted that other comment that pointed out this issue. So I will ask you again, why did you do it? What are you so afraid of?
@ShahidkBaloch
@ShahidkBaloch 2 ай бұрын
Good 👍
@galahad-
@galahad- 2 ай бұрын
as usual, juste bloody so interesting
@nagashud
@nagashud 2 ай бұрын
I usually use Inworld on my phone, and by the explorer it uses so much phone battery and memory, so is it possible in the future to have it as an app ?
@inworldai
@inworldai 2 ай бұрын
It's not currently in our product roadmap as we're focused mostly on B2B use cases. However, hopefully someone will integrate our tech into an app that you can use!
@nagashud
@nagashud 2 ай бұрын
@@inworldai thank you for your answer! That would be awesome !
@renejersild6825
@renejersild6825 2 ай бұрын
I want to make a NPC model, based on a nonverbal autistic person 30 years old, cognitive 3 years old. Its for a learning enviroment. Is that possible with Inworld?
@inworldai
@inworldai 2 ай бұрын
Yes! We have been experimenting with powering gameplay for non-speaking characters and objects. If you go to our Channel's homepage there's an AI coop video showing non-verbal interactions. We also, internally, did a demo recently with animals. For more info about how to do this, I recommend joining our Discord and asking our community. We don't monitor comments on KZbin frequently.
@BenEncounters
@BenEncounters 2 ай бұрын
If I understand well, you can trigger a new scene through Goals & Actions and each new scene will have different environments, as well as different characters in them, but also each character might be in a different state of mind with new motivations? And when a new scene is triggered, how does it look like of feel like in the basic Chatbot view?
@inworldai
@inworldai 2 ай бұрын
Our chatbot functionality is still in the process of adding multi-agent conversations to it so you can currently only speak to one character at a time. But that feature is launching soon. I recommend joining our Discord to ask our community and team questions there!
2 ай бұрын
Hola AngeliK, cómo estás? Quería hacerte una consulta, se puede usar una ia personalizada o solo se puede conectar con la ia de ustedes? Muchas gracias
@inworldai
@inworldai 2 ай бұрын
Hola Cristian, estoy bien, ¡gracias por preguntar! Si entendí bien, ¿cuando hablas de personalizar la IA te refieres al físico del personaje? Si es así, solo tendrías que conectar el Brain ID y ponerle los componentes necesarios en Unity. Si te refieres a poder personalizar la IA como tal, todas las funciones disponibles para personalizar están en Inworld Studio.
2 ай бұрын
@@inworldai gracias por responder 🥰 Si fuese posible hacer las dos cosas genial! Porque en realidad tengo un avatar propio que desearía utilizar en vez de tener que utilizar uno desde Playerme. Y por otro lado, mi deseo es utilizar una IA entrenada con documentación específica y cuyas respuestas queden grabadas de una base de datos personal. Serían realizables estas propuestas? Muchísimas gracias de nuevo y es excelente tu onda en los videos tutoriales 👍
@annagayete1011
@annagayete1011 2 ай бұрын
My character is not talking and I was following everything here.
@inworldai
@inworldai 2 ай бұрын
If you're still having difficulties, I recommend joining Discord as we provide support through that channel. We're happy to answer your questions there.
@irfanmustaqim948
@irfanmustaqim948 2 ай бұрын
Can this features been used to trigger some animations that will be played by the character ? How do I set it ?
@inworldai
@inworldai 2 ай бұрын
Yes! I recommend joining Discord as we provide support through that channel. We're happy to answer your questions there.
@irfanmustaqim948
@irfanmustaqim948 2 ай бұрын
Great video ! Can I add custom animation to the character to perform as I said something to trigger the animation ? As example, when I said "Please Dance" to the character and the character will respond to it by playing the "Dance" animation.
@inworldai
@inworldai 2 ай бұрын
Yes, you can do that through Goals by using training phrases and then sending a trigger notification to Unity. You will need to create a script that receives the trigger and plays the "Dance" animation. The next video will explain Goals, so stay tuned!
@zaarbeats
@zaarbeats 2 ай бұрын
Can we have a tutorial on how to release game?
@inworldai
@inworldai 2 ай бұрын
Wrote it for the pipeline :D !
@alystervodka8799
@alystervodka8799 2 ай бұрын
well.. i followed all the inworld AI tutorials but none of them explain how to set up the OVR lip sync playback actor, and it didn't come with the Inworld plugin.
@inworldai
@inworldai 2 ай бұрын
If you're still having difficulties, I recommend joining Discord as we provide support through that channel. We're happy to answer your questions there. I believe it is now incorporated in our SDK but they'll be able to help you there best.
@SivaneMarciano
@SivaneMarciano 2 ай бұрын
Thank you so so much for those tutorials !! Can you please make a video on how to create a widget to interact with the character with text, on Unreal, as show cased in the playground? I couldn't find documentation on this... Thanks again !!😊
@inworldai
@inworldai 2 ай бұрын
Writing it into the pipeline! Thank you <3
@adriangarces91
@adriangarces91 2 ай бұрын
Hello, thank you very much for this incredible IA tool and for all the tutorials. But when will they have a Spanish language?
@inworldai
@inworldai 2 ай бұрын
Hola Adrian! Por ahora, solo estamos ofreciendo inglés, mandarín y coreano. Para las próximas expansiones tenemos el español en el listado, pero no tenemos una fecha aún. Sin embargo, muchos de nuestros usuarios han creado soluciones temporeras para que los personajes puedan hablar español, así que te recomiendo que te unas a nuestro Discord y veas sus ejemplos!
@SajiVlogsSL
@SajiVlogsSL 2 ай бұрын
Im new to this tutorial. Does the 8th wall have a project already for us to clone ? How can i start my own new idea and build this up ? Please give some guidance . Thank you !!!
@inworldai
@inworldai 2 ай бұрын
Yes. We do have a project in 8th Wall you can clone. If you have any other questions, we recommend joining our discord because we answer support requests and questions there!
@shiv-et6if
@shiv-et6if 2 ай бұрын
The proximity thing adds so much realism to this. Excited to play with this
@lezerpen
@lezerpen 2 ай бұрын
From what I’ve read is this based on GPT3. Are there any plans to use a newer version of GPT?
@inworldai
@inworldai 2 ай бұрын
We actually dynamically switch between multiple commercial and proprietary or open source models based on which is the best for the conversational context and which will return the response fastest based on each model's current real-time latency. I believe we used GPT-4 for some things but the latency was too high for most real-time uses. GPT-4o is much faster so much more possible to add to the mix. We also provide fine tuning to clients who want more specialized use cases. For more information about how we benchmark and measure conversational quality -- which helps us choose which model to use for which kinds of queries --, check out our blog for a paper we just put out on the subject. .
@jrichardsio
@jrichardsio 2 ай бұрын
This was great, but you should show the final example ha
@inworldai
@inworldai 2 ай бұрын
Ok! Will keep that in mind 🔥
@Showbiz_CH
@Showbiz_CH 2 ай бұрын
Do the NPCs talk to each other?
@inworldai
@inworldai 2 ай бұрын
We are launching our multi-agent feature soon that will allow them to but in this mod, they do not.
@Showbiz_CH
@Showbiz_CH 2 ай бұрын
@@inworldai thanks! This will be interesting!
@Conbini.
@Conbini. 2 ай бұрын
Was there a part where you describe how to install Unreal from source code on a Mac? I just installed from the Epic Games launcher and it does not install a source code version by default... as in, there's no .SLN files that get created every time you make a Uproject...
@inworldai
@inworldai 2 ай бұрын
That process is described in 01:25 and if you're doing a non-marketplace build in our plugin then you have to create a code project. However, if you are having more questions, then reach out to us at [email protected] to tackle your case more specifically.
@Conbini.
@Conbini. 2 ай бұрын
@@inworldai Thanks, so in other words I can do this same thing with regular Unreal using the marketplace plugin, right?
@inworldai
@inworldai 2 ай бұрын
@@Conbini. Yes you can!
@dyna4studio942
@dyna4studio942 2 ай бұрын
Love how this girl teaches!
@samayres4108
@samayres4108 3 ай бұрын
It would be quite interesting if the NPCs using inworld responded in Portuguese
@inworldai
@inworldai 2 ай бұрын
We don't offer official support yet, but if you visit our Discord, you'll see ingenious ways other developers have come up with to find temporary solutions. 👀
@samayres4108
@samayres4108 3 ай бұрын
I'm loving the tutorials, they've helped me a lot, I would really like there to be a Portuguese-Brazilian translation of the lines.
@inworldai
@inworldai 2 ай бұрын
Taking notes, let me see what I can do !
@waqaskhan-uw3pf
@waqaskhan-uw3pf 3 ай бұрын
Hi i tried to import my ready me prefab but it didn't work. The integration went well. Access token connected and the scene with my newly created character shown in the box but in assets folder the imported prefab is the default basic innequin . the thumbnail in the inworld panel is of my created character but the prefab 3d chrctr is not the one that i created. Please guide me to fix
@inworldai
@inworldai 2 ай бұрын
We can definitely help you with this! Email us at [email protected], and we can provide you with more personalized assistance.