NNiceVois
Focused training replacement

A Weights AI alternative for training your own RVC model

Train my voiceKept your old .pth from Weights? Use it in the voice converter.
Summary

NiceVois covers the part of Weights.gg you can own: train an RVC model from your own voice, download the .pth and .index, and use it to convert vocals or whole songs. A .pth you kept from Weights works too. What it does not have is a public catalogue of other people’s voices, and that is on purpose.

Weights AI — the service most people knew as Weights.gg — closed on 1 April 2026. If the part you miss is getting a portable RVC model without setting up local training or repairing a notebook, this utility is designed for that exact job.

Voice training

Train an RVC voice from an audio file

Add up to 50 WAV, MP3, FLAC, M4A, or OGG files. NiceVois keeps them separate and prepares them together as one private training dataset.

.pth

This name is used for the .pth, .index, and complete ZIP package.

Standard is selected automatically.
200 epochs

Recommended: add your audio and NiceVois will suggest the right number.

Clean audio produces a better model. Use a dry voice recording with minimal music, echo, or background noise. Prepare your audio first if it needs it.

Persistent cloud storage

Your model library

Your voice models are files you own. Download them once and no platform can take them away.

Loading…
Paid model packages stay saved until you delete them.

Your .pth, .index, ZIP, and private listening preview remain available after closing the browser or restarting your computer. A model from the free training run is kept for 7 days and shows its expiry date; download it before then. Full-resolution source audio and temporary training files are removed after validation.

Loading your models

Checking private cloud storage…

What this alternative provides

The free tools around the trainer go further than the ones you are replacing: besides vocals and instrumental, NiceVois can split the lead from the background vocals, so an AI cover keeps the original harmonies instead of converting them.

What it does and does not replace

CapabilityNiceVois
Train your own RVC voice modelYes
Download portable model filesYes: .pth, .index, and ZIP
Convert an isolated vocalYes: normal RVC and NiceVois-cleaned outputs
Use a .pth you exported from WeightsYes: upload it to the voice converter
Create an AI cover inside the siteYes: turn a finished song into your voice
Browse a public model catalogNo
Publish or remix another person’s modelNo

Why portable files matter

A downloadable model is not locked to one hosted conversion interface. The .pth contains the trained RVC weights, and the matching .index supports retrieval of speaker-specific features in compatible tools. Keeping the complete ZIP gives you an independent copy that can outlive the training service.

That is not a hypothetical. When Weights.gg closed, the people who lost nothing were the ones who had exported their files. The case study goes through what the platform got right and why the free model could not last.

What you need before training

Use your own voice or obtain clear permission from the speaker. Prepare a clean, single-speaker recording with minimal music, echo, clipping, or competing voices. The first training is free with an account, for up to 5 minutes of audio at up to 150 epochs; paid trainings take longer recordings and up to 500 epochs, and every training shows its price before it starts. A model from the free training is kept for 7 days, so download it; models from paid trainings stay until you delete them.

Rebuild what you had on Weights

Train your voice and leave with portable RVC files, or bring the .pth you kept and use it in the converter.

Train my voice