male_singing_1

male_singing_2

female_singing_1

female_singing_2

Unofficial fan-made AI voice model of TREE, an early joke UTAU character originally created by annamaeblythe / Winchester24 and voiced by Winchester24. TREE is best known for his harsh, growling, scream-like vocal samples and intentionally chaotic singing style. This reconstruction is based on recovered original TREE voicebank material, with the goal of preserving his abrasive vocal character rather than smoothing it into a conventional synth voice. Selected surviving performance references may also be used for additional phrasing and expression. This is an unofficial fan preservation project and is not an official release, replacement, or continuation of the original UTAU voicebank.


Credits go to meaisae on TikTok (not me ofc)

Unofficial fan-made AI voice model of Karomori (カロの森), a male anthropomorphic UTAU/Furloid created by MeoRoo. Karomori is characterized as gentle, selfless, and timid, and his Japanese VCV voicebank was designed for smoother, more natural transitions between syllables. This model uses archived Karomori voicebank material plus selected surviving song references to capture more of his phrasing and expression. This is not an official release or replacement for the original UTAU voicebank.

Unofficial fan-made AI voice model of Beat Okamine (狼音ビート), a 19-year-old male wolf UTAU/Furloid created and voiced by Missa-kun, also known as AlexVocaloid/Ritmos. Originally released in 2012, Beat was the creator’s first UTAU project and has been developed through multiple voicebank formats and updates over the years, including CV, VCV, multilingual banks, and expressive appends. This reconstruction uses Beat’s archived VCV ACT 1 voicebank as its core dataset, preserving his original vocal identity while selected surviving song recordings and WILD/Appends demonstrations are kept separately as supplemental references for phrasing, sustained singing, tone, and expression. Beat’s voice is especially suited to J-pop, J-rock, electronic music, and energetic or emotional material. This is an unofficial preservation/fan reconstruction project and is not an official release, replacement, or continuation of the original UTAU voicebank.



Taka Ototsuki / 音憑タカ is a male UTAU voiced by Yuuri and originally released in 2012. Although technically human in his lore, Taka can transform into animals and animal-like forms. His Beast form combines fox and avian traits, giving him one of the more unusual designs among older UTAUs. This model uses recordings from his Normal, Beast, Soft, and High voicebanks to preserve both his core tone and expressive range.

Competitor on the show American Ninja Warrior (prototype voice)

Unofficial fan-made AI voice model of SKIPLOID, an UTAU/Furloid originally created by Pokeluver223. Trained from archived SKIPLOID voicebank material and surviving song demos/covers, including fix, Close To You, ODDS & ENDS, and Reversible Campaign. SKIPLOID has a soft male vocal tone suited to pop, ballads, electronic music, and emotional songs. This model is not an official release or replacement for the original UTAU voicebank.

Use this for Sora:
Sora is a male UTAU created by RyukaK in 2013. He is a 22-year-old four-tailed lion with a distinctly old-school UTAU sound. His original Japanese CV voicebank has a warm, straightforward tone that fits melodic, emotional, and classic UTAU-style songs especially well.
This model was built from Sora’s preserved original voicebank recordings and is intended to retain the character and tone of the original library.

Experimental fan reconstruction trained on archived UTAU voicebank samples attributed to 信幸大和猫 / Yamato. This model is for preservation and testing purposes and is not an official release or replacement for the original voicebank.


Unofficial fan-made AI voice model of Kuroda Samene (黒田 鮫音), a 25-year-old male shark UTAU/Furloid created by Missa-kun, also known as AlexVocaloid/Ritmos. Kuroda is the creator’s second UTAU project after Beat Okamine and is known for his distinctly deep, masculine vocal character. His original voice was created using alternate, deeper-recorded samples that were then pitched down to achieve Kuroda’s signature low, heavy vocal tone. This reconstruction uses two archived Kuroda voicebanks, including VCV ACT 2 and an earlier bank, giving the model a broader sample of his original vocal identity across different recording sets. Selected surviving song and medley recordings are kept separately as supplemental references for phrasing, sustained singing, and expression. This is an unofficial preservation/fan reconstruction project and is not an official release, replacement, or continuation of the original UTAU voicebank.

GiannaCapizzi