







Yattaaa~!
Today we go genki π then some gaming after roleplaying β€οΈ
Selfies of today β€οΈ








Today we go genki π then some gaming after roleplaying β€οΈ
Selfies of today β€οΈ
Hii all, this is Kim. I opened my art stream on April this year, my goal is to create a nice and friendly community. I started this new segment of my VTube channel, today: Puri, the roaster.
Today's selfies! π
Credits to Airi on the avi π₯
Follow me on Twitch :)Β β€οΈ
Hii all, this is Kim. I opened my art stream on April this year, my goal is to create a nice and friendly community. I started this new segment of my VTube channel, today: Puri, the roaster.
Today's selfies! π
Credits to Airi on the avi π₯
Follow me on Twitch :)Β β€οΈ
Hii all, this is Kim. I opened my art stream on April this year, my goal is to create a nice and friendly community. I started this new segment of my VTube channel, today: Puri, the roaster.
Today's selfies! π
Credits to Airi on the avi π₯
Follow me on Twitch :)Β β€οΈ
Hii all, this is Kim. I opened my art stream on April this year, my goal is to create a nice and friendly community. I started this new segment of my VTube channel, today: Puri, the roaster.
Today's selfies! π
Credits to Airi on the avi π₯
Follow me on Twitch :)Β β€οΈ
Hi guys, I'm Kim, I started streaming on April this year. I started streaming fan fiction stories, I mean, BGM, SFX, voices, all those things I did from scratch to create every story. Now I'm aiming towards character roleplaying, today I performed as Frieren the whole day on VRC. Basically the question is about where to stream my content, I mean, I mostly stream on Twitch but I feel this platform is more for gaming? what do you think?
Hi guys, I'm Kim, I started streaming since April, we are aiming to build a nice and friendly community, wanna join us?
Today's selfie!
Hi guys, I'm Kim, I started streaming since April, we are aiming to build a nice and friendly community, wanna join us?
Comedy full of memes! π
Today's selfie! β€οΈ
I love you guys! I'll be waiting for you β€οΈ
We aim to build a friendly community, everyone is welcome π
Expect the unexpected! π
Comedy full of memes! π
Today's selfie! β€οΈ oops! I blinked there! π
I love you guys! I'll be waiting for you β€οΈ
I wish you a lovely day β€οΈ
Thanks for your attention. I wish you a lovely day β€οΈ
Expect the unexpected π
Infinite comedy! π
Today's selfie! β€οΈ
I love you guys! I'll be waiting for you β€οΈ
https://twitch.tv/laguprincess
Laguprincess is a 24/7 Twitch streamer that plays VRChat. It's currently live on twitch.tv/laguprincess
She can perform something I call "world hop". When the profanity level is too high or there are no interactions in N time she chooses to perform a world hop to hop into a new public lobby, the most populated world wins. Then VRC automatically reopens in the new selected world. Once there, she will always go towards the speech, she has multiple walking states, such as: approach, socialize, wandering. In the first one, she already heard something, she knows exactly where the sound is coming from and aims her direction towards the speaker, once she gets into the socializing radius, she swaps to socialize state. In socialize state, she doesn't move, she always do 'facing' she will only rotate to face the speaker. In wandering state, she's not hearing anything, she randomly moves. In socializing state she can remember N conversation turns or how long I want her to remember the conversation history.
The way she hears is through voice audio detection, then passed to speech to text, then some LLM probe decision making like "is this worth answering to?", generate LLM response using ResponseBuilder(). ResponseBuilder is also responsible to decide whether the response will be spoken by what TTS engine (provided by Ontology API). She is always in high quality mode which means I can send a timbre_reference (the actual voice identifier) .wav file and style_reference (singing, accents, languages, rapping, sing opera, whatever) .wav file for the inference. She can speak in English, German, Japanese, Chinese, French, Hindi. Switch language based on what she hears. The low quality mode supports English (US), English (US), Spanish, French, Hindi, Italian, Japanese, Portuguese (BR), Chinese. This is a real-time speed TTS, here you just select what voice on what language, the speed you want it to be spoken.
What she hears is also animated through an OBS browser source, using HTML5 + JS. 20ms per character.
What she speaks is animated through another OBS browser source, this one goes in three phases:
The first phase is the wandering phase: N configured colored particles wandering around while the response is processing.
The second phase is the forming phase: Reuse the N configured colored wandering particles to form the final response character by character. The result is a cute rainbow font.
The third phase is the dancing phase: Every character is dancing at its own pace very smoothly.
She is connected to Twitch using IRC. She is aware of online viewers, subscribers, she has viewer notes that are later attached into the prompt to generate more personalized responses for the twitch chat. She reacts on follows, subscriptions, Viewer rewards.
She have several custom viewer rewards that activate in-game and/or voice actions. The viewers can make her jump, send a sanitized text to VRChat, make her scream, and say poi poi poi.
Hi guys, I want to build a cute and nice community in my stream, everybody is welcome! π
Expect the unexpected π
Infinite comedy! π
Today's selfie! β€οΈ
I love you guys! I'll be waiting for you β€οΈ
We aim to build a friendly community, everyone is welcome π
Thanks for your attention. I wish you a lovely day β€οΈ
Hi guys, I play mostly VRChat, wanna come say hi?