Voices for AI covers
Pick a community voice model and make an AI cover of any song, or bring your own.
Popular voices
SpongeBob SquarePantsCartoon character
50 CentRapper
EminemRapper
DrakeRapper
CaseOhYouTuber and streamer
SpongeBobSpongeBob SquarePants
Michael JacksonKing of Pop
Taylor SwiftSinger
Ariana GrandePop singer
Arnold SchwarzeneggerActor
Kanye WestRapper
Joe RoganThe Joe Rogan Experience
PlanktonSpongeBob SquarePants
BlueyBluey
Juice WRLDRapper
Justin BieberPop singer
SonicSonic the Hedgehog
Bad BunnyReggaeton singer
KATSEYEGirl group
Sandy CheeksSpongeBob SquarePants
MadonnaPop singer
Hatsune MikuVocaloid
SquidwardSpongeBob SquarePants
Freddie MercuryQueen
Homer SimpsonThe Simpsons
Twilight SparkleMy Little Pony
Atif AslamPakistani singer
Cristiano RonaldoFootballer
Kasane TetoVocaloid
Mr. KrabsSpongeBob SquarePants
Neco ArcMelty Blood
Peter GriffinFamily Guy
RihannaSinger
Billie Eilishsinger
King VonRapper
Lana Del ReySinger
Playboi CartiRapper
Sidhu Moose WalaPunjabi singer
The WeekndSinger
Bruno MarsPop singer
Chester BenningtonLinkin Park
Frank SinatraSinger
Kurt CobainNirvana
MarioSuper Mario
MegadethThrash metal band
Mickey MouseDisney
PomniThe Amazing Digital Circus
Elvis PresleyRock and roll singer
Eric CartmanSouth Park
Lady GagaPop singer
Nicki MinajRapper
Patrick StarSpongeBob SquarePants
Gawr GuraHololive
Rainbow DashMy Little Pony
MrBeastYouTuber
Natanael CanoMexican rapper
Peso PlumaMexican singer
Light YagamiDeath Note
Luis MiguelSpanish-language singer
Master ChiefHalo
About RVC voice models
Most AI cover voices are RVC (Retrieval-based Voice Conversion) models. Each one is trained on about 10 to 50 minutes of clean recordings of a single voice, and can then sing any song in that voice while keeping the original melody and timing.
If you saved RVC models from Weights.gg or elsewhere, you can import them here. You can also train a new model of your own voice from 10 minutes or more of a cappella singing.
Questions
What is a voice model?
A small AI model trained on recordings of one voice. Our covers use RVC models: a .pth file with the trained voice and an .index file that helps it sound like the original.
Can I use my own RVC models?
Yes. Import the .pth and .index files of an RVC v2 model and keep it private or share it with the community.
Who makes these voices?
Users share them, and we also import models from public community collections such as Hugging Face, credited to the Weights AI Library account. Voices of real people are unofficial fan models. Anyone can report a model, and we review reports within 24 hours.