awesome-genmedia Rvc v2
Convert voices using RVC (Retrieval-based Voice Conversion) v2. Apply pre-trained or custom voice models to audio with fine-grained control over pitch, accent blending, and output quality.
Quick Start
Requires an each::labs API key. Get one at eachlabs.ai.
curl -X POST https://api.eachlabs.ai/v1/prediction \
-H "Content-Type: application/json" \
-H "X-API-Key: $EACHLABS_API_KEY" \
-d '{
"model": "rvc-v2",
"version": "0.0.1",
"input": {
"input_audio": "https://example.com/original-vocals.wav",
"rvc_model": "Obama",
"f0_method": "rmvpe",
"pitch_change": 0,
"output_format": "wav"
}
}'
Parameters
| Parameter | Type | Default | Description |
|---|---|---|---|
| crepe_hop_length | integer | 128 | When f0_method is mangio-crepe, controls how often it checks for pitch changes. |
| custom_rvc_model_download_url | string | - | URL to download a custom RVC model. If provided, the model will be downloaded. |
| f0_method | string | rmvpe | Pitch detection method. Options: rmvpe, mangio-crepe |
| filter_radius | integer | 3 | If >=3, apply median filtering to the harvested pitch results. |
| index_rate | number | 0.5 | Control how much of the AI's accent to leave in the vocals. |
| input_audio | string | - | URL of the audio file to convert. |
| output_format | string | wav | Output format. Options: mp3, wav |
| pitch_change | number | 0 | Adjust pitch of AI vocals in semitones. Positive to increase, negative to decrease. |
| protect | number | 0.33 | Control how much of the original vocals' breath and voiceless consonants to leave. |
| rms_mix_rate | number | 0.25 | Control how much to use the original vocal's loudness (0) or a fixed loudness (1). |
| rvc_model | string | CUSTOM | Voice model to use. Options: Obama, Trump, Sandy, Rogan, CUSTOM |
Examples
Using a built-in model:
curl -X POST https://api.eachlabs.ai/v1/prediction \
-H "Content-Type: application/json" \
-H "X-API-Key: $EACHLABS_API_KEY" \
-d '{
"model": "rvc-v2",
"version": "0.0.1",
"input": {
"input_audio": "https://example.com/singing.wav",
"rvc_model": "Trump",
"pitch_change": -2,
"output_format": "mp3"
}
}'
Using a custom model with fine-tuned settings:
curl -X POST https://api.eachlabs.ai/v1/prediction \
-H "Content-Type: application/json" \
-H "X-API-Key: $EACHLABS_API_KEY" \
-d '{
"model": "rvc-v2",
"version": "0.0.1",
"input": {
"input_audio": "https://example.com/vocals.wav",
"rvc_model": "CUSTOM",
"custom_rvc_model_download_url": "https://example.com/my-trained-model.zip",
"f0_method": "rmvpe",
"index_rate": 0.7,
"pitch_change": 3,
"protect": 0.4,
"rms_mix_rate": 0.3,
"output_format": "wav"
}
}'
Related Models
- Train Rvc - Train a custom RVC model from audio datasets
- Realistic Voice Cloning - Clone voices from audio samples
- ElevenLabs | Voice Changer - ElevenLabs voice changing