# What is RVC? Retrieval-based Voice Conversion explained

> RVC (Retrieval-based Voice Conversion) is open-source AI that re-sings or re-speaks a recording in another voice, keeping the words, melody and timing. An RVC model is a .pth file, usually with an .index file, trained on one voice. Use RVC voices in the browser on Jammable, upload your own model files, or run them locally with Applio.

- RVC changes who is singing, not what is sung: the melody, words and timing come from your input.
- The .pth file is the voice itself; the .index file is a lookup table of the voice's sounds that makes the result sound more like the target.
- RVC is not text-to-speech: it needs a recording to convert. Jammable's text-to-speech speaks your text first, then converts it.
- Pitch is the setting that matters most: +12 or -12 semitones when the input and the voice sit an octave apart.
- Clean, dry vocals in and a well-trained model are what make a good result.

Source: https://www.jammable.com/blog/what-is-rvc  
Published: 2026-10-17  
Updated: 2026-10-17  
Publisher: Jammable

## What is RVC?

RVC stands for **Retrieval-based Voice Conversion**. It is a free, open-source voice conversion framework published on GitHub by the RVC-Project team under the MIT license, built around a web app (the RVC WebUI) for training and using voice models. Its own tagline is "easily train a good VC model with voice data <= 10 mins".

Voice conversion means taking a recording of one voice and making it sound like another. RVC keeps what is being said or sung (the words, the melody, the timing, the phrasing) and swaps the voice that says it. That is why it became the standard way to make AI covers: you give it a vocal, it gives you the same performance in a different voice.

An **RVC model** is what you get when you train RVC on recordings of one voice. Community voice libraries are mostly RVC models, and Jammable trains RVC models too, so a voice you [train on Jammable](https://www.jammable.com/blog/how-to-train-an-ai-voice-model) is the same kind of model people run on their own computers.

## What does "retrieval-based" mean?

Most voice conversion has a problem called timbre leakage: a little of the original singer's tone survives the conversion, so the result sounds like a blend of two people. RVC's fix is in its name. When it trains a voice, it also saves the sound features of the training audio in a searchable index. When it converts, it looks up the closest matches in that index for each moment of your input and blends them in, so the output is built from sounds the target voice actually made.

The RVC README describes this as using retrieval "to replace the source feature to training-set feature" to reduce tone leakage. How strongly it does that is a setting, usually called the index rate or feature ratio.

## How RVC turns one voice into another

1. **Separate the vocal.** RVC works on a voice on its own. For a full song, the vocal is first split from the music. Jammable does this step for you; locally you use a separation tool first.
2. **Read the content.** A speech model (ContentVec or a HuBERT model) turns the vocal into features that describe what is being said or sung, with as little of the original voice in them as possible.
3. **Track the pitch.** A pitch extractor follows the melody note by note. RVC's recommended extractor is RMVPE; Applio also offers crepe, fcpe and swift. Transpose shifts this pitch curve up or down.
4. **Look up the target voice.** If there is an .index file, RVC finds the closest matching features from the training audio and blends them in, by as much as the index rate says.
5. **Generate the new voice.** The .pth model turns the content and the pitch into audio in the target voice. For a cover, that vocal is then mixed back over the music.

## The .pth and .index files

What is in an RVC model folder, based on the RVC FAQ and the RVC WebUI

| File | What it is | Do you need it? |
| --- | --- | --- |
| .pth (the model) | The trained voice: the network that generates audio in that voice. The RVC FAQ calls it the 60+ MB file to share. | Yes. Without it there is no voice. |
| .index (the feature index) | A searchable table of the training audio's sound features, used for retrieval. Built during training. | Optional. RVC runs without it, but most voices sound closer to the target with it. |
| G_ and D_ checkpoints | Training snapshots kept while a model trains, used to resume training. | No. The RVC FAQ says they are not for sharing or inference. |

The RVC FAQ says that with long, high-quality training audio the index matters less, and some people skip it. With shorter or noisier datasets it is what stops the result drifting back towards the original singer. If you have both files for a voice, keep them together.

## What is RVC v2?

RVC v2 is the second generation of RVC's pretrained base models, the starting point every new voice is trained from. The RVC changelog added v2 base models in May 2023 (a 40k sample-rate version first, then 32k and 48k in June 2023). A v2 model uses the content encoder's full 768-value features for each moment of audio, where v1 squeezed them down to 256, and Applio trains v2 models by default.

You usually don't need to do anything about it: Applio reads the version from the model file and runs v1 and v2 models alike. The RVC README says a v3 base model, larger and trained on more data, is on the way.

## RVC vs text-to-speech vs real-time voice changers

Three kinds of voice AI that are often confused. RVC's real-time mode is its separate desktop GUI.

| Feature | RVC voice conversion | Text-to-speech | Real-time voice changer |
| --- | --- | --- | --- |
| Starts from | A recording | Typed text | Your live microphone |
| Keeps your melody and timing | Yes | No | Yes |
| Can sing | Yes | No | Partly |
| Works live on calls and streams | Partly | No | Yes |
| Best for | Covers and re-voicing | Narration and voiceovers | Gaming and streaming |

**Text-to-speech** creates speech from text, so it decides the rhythm and delivery itself. **RVC** needs a performance to convert, which is why it can sing: the singing comes from the input. **Real-time voice changers** convert your microphone as you speak. RVC can do that too: its WebUI ships a real-time GUI, and the README reports about 170 ms end-to-end latency (90 ms with ASIO audio devices, depending on drivers). Applio has a real-time tab as well.

The lines blur in practice. Jammable's text-to-speech speaks your text in a base voice first and then converts it with the RVC voice you picked, so you can hear typed text in a community voice. Jammable does not change your voice live, so for calls or streams you need a desktop tool. For converting a recording you made, see the [AI voice changer](https://www.jammable.com/ai-voice-changer).

## How to use an RVC model

There are three ways, depending on whether you have model files and how much setup you want.

- **Pick a voice in the browser.** Jammable's [voice library](https://www.jammable.com/models) has more than 60,000 community voices, all ready to use. Nothing to install, and it works on a phone. Covers need a plan, from $1.99 for the first month.
- **Upload your own .pth and .index.** If you trained a model elsewhere or kept files from another site, upload them to Jammable and use the voice like any other.
- **Run it locally.** The RVC WebUI and Applio are free, open source and run on your own computer. RVC's README covers NVIDIA graphics cards as well as AMD, Intel and CPU-only machines. You control every setting, and you find and manage model files yourself.

## How to upload a .pth and .index file to Jammable

1. **Open your Library.** Go to your [Library](https://www.jammable.com/library) and choose *Upload a voice*.
2. **Add the files.** Drop in the .pth file (required) and its .index file (optional, but recommended). Jammable checks the files and refuses ones that are invalid or corrupt.
3. **Name it and add a picture.** Give the voice a name of up to 40 characters, a picture and, if you like, a short description and categories.
4. **Choose who can use it.** Pick *Private, only me* or *Public in the library*, agree to the Terms of Use and press Publish. The voice gets its own page and is ready for covers and text-to-speech.

Only upload models of your own voice, or of a voice whose owner has agreed to it.

## How to get good results with RVC

- **Set the pitch first.** Pitch (called transpose in RVC) is in semitones, and 12 semitones is one octave. If the input is a lower voice and the model is a higher one, try +12; the other way round, try -12. On Jammable the range is -24 to +24, and on a single-voice cover a shift that isn't a whole octave moves the music by the same amount so the vocal stays in key.
- **Feed it a clean vocal.** RVC converts whatever it hears, including reverb, harmonies and background noise. On Jammable, turn on *Split out backing vocals* when a song has harmonies or ad-libs, and *My audio is just vocals* when your file is already a dry vocal.
- **Use the index, but not always at full strength.** Locally, the index rate (Applio calls it *Search Feature Ratio*, default 0.75) trades closeness to the target voice against artifacts. Lower it if the result sounds metallic.
- **Protect breaths and consonants.** RVC's *protect* setting keeps "s", "t" and breath sounds from tearing. In the RVC WebUI, 0.5 turns it off and lower values protect more, at some cost to the index's effect.
- **Use RMVPE for pitch.** It is the default in Applio and the extractor RVC recommends: fast, light and good at avoiding dropped notes.
- **Start from a good model.** No setting rescues a voice trained on noisy audio. The RVC FAQ recommends 10 to 50 minutes of clean audio (5 to 10 minutes can be enough if it is very clean). Jammable accepts 3 minutes and voices usually sound better with 6 or more.

> **Robotic or wobbly result? Check the octave** A voice pushed outside its range is a common cause of a bad conversion. If the result sounds strained, thin or full of cracks, try the pitch 12 semitones the other way before changing anything else.

[Browse RVC voices](https://www.jammable.com/models): Pick one of 60,000+ community voices and use it in your browser.

## FAQ

### What is an RVC model?

An RVC model is a voice trained with Retrieval-based Voice Conversion. It is a .pth file, usually with an .index file, that can re-sing or re-speak a recording in that voice. Most community AI voices, including the ones on Jammable, are RVC models.

### What is the difference between the .pth and the .index file?

The .pth file is the trained voice itself. The .index file is a searchable table of the sounds in the training audio, which RVC uses to make the result sound closer to the target voice. You need the .pth; the .index is optional but usually helps.

### How do I use an RVC model?

Pick a voice in Jammable's library and use it in your browser, upload your own .pth and .index to Jammable, or run the model locally with the free RVC WebUI or Applio on your own computer.

### Is RVC a voice changer?

RVC is a voice conversion framework. It converts recordings, and its desktop GUI can also change your voice live with roughly 170 ms of latency. Jammable uses RVC voices for recorded audio and text-to-speech, not for live calls or streams.

### Is RVC free?

Yes. RVC and Applio are open source under the MIT license and free to run on your own computer. Using RVC voices on Jammable needs a plan, from $1.99 for the first month, but you can browse the voices and listen to demos for free.

### What is RVC v2?

RVC v2 is the second generation of RVC's pretrained base models, added in 2023. v2 models use richer content features than v1. Applio loads both versions and trains v2 by default, so you rarely need to worry about which one a model is.

### How much audio does it take to train an RVC model?

RVC's own FAQ recommends 10 to 50 minutes of clean audio, or 5 to 10 minutes if it is very high quality. Jammable accepts 3 minutes of singing or speech and voices usually sound better with 6 minutes or more.

## Related

- [Browse 60,000+ AI voices](https://www.jammable.com/models)
- [AI voice models explained](https://www.jammable.com/blog/ai-voice-models)
- [Train your own voice](https://www.jammable.com/blog/how-to-train-an-ai-voice-model)
- [Weights AI alternative](https://www.jammable.com/weights-gg-alternative)
- [Pricing](https://www.jammable.com/pricing)

## Sources

- [RVC-Project on GitHub](https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI)
- [RVC FAQ (English)](https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/blob/main/docs/en/faq_en.md)
- [RVC changelog (English)](https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/blob/main/docs/en/Changelog_EN.md)
- [Applio on GitHub](https://github.com/IAHispano/Applio)
- [Applio inference settings](https://github.com/IAHispano/Applio/blob/main/tabs/inference/inference.py)
- [Applio training settings](https://github.com/IAHispano/Applio/blob/main/tabs/train/train.py)