> ## Documentation Index
> Fetch the complete documentation index at: https://docs.youka.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Sync Model Reference

> Pick the right lyric synchronization model for your song

Youka has two kinds of sync models: alignment models that use lyrics you provide, and transcription models that detect lyrics from the audio.

## Model comparison

| Model                          | Type          | Provide lyrics? | Best for                                             |
| ------------------------------ | ------------- | --------------- | ---------------------------------------------------- |
| Audioshake Alignment (default) | Alignment     | Yes             | English pop/rock with accurate lyrics                |
| MusicAI Alignment              | Alignment     | Yes             | When Audioshake Alignment gives poor timing          |
| Audioshake Transcription       | Transcription | No              | English songs when you don't have lyrics             |
| MusicAI Transcription          | Transcription | No              | Non-English songs when you don't have lyrics         |
| Whisper AI                     | Transcription | No              | Multilingual songs or non-English speech recognition |

## Quick picker

| Your song                     | Recommended model                                          |
| ----------------------------- | ---------------------------------------------------------- |
| English pop, have lyrics      | Audioshake Alignment                                       |
| Unknown or missing lyrics     | Audioshake Transcription                                   |
| Non-English song              | Whisper AI or MusicAI Transcription                        |
| Tricky timing after auto-sync | Any model, then [Manual Sync](/en/desktop/fix-lyrics-sync) |

## How to pick a model

<Steps>
  <Step title="Open the lyrics editor">
    Go to **Edit > Edit Lyrics**.
  </Step>

  <Step title="Enter lyrics (for alignment models)">
    Paste or edit the lyrics if you're using an alignment model. Leave the field blank for transcription models.
  </Step>

  <Step title="Select the model">
    Pick your preferred sync model from the dropdown.
  </Step>

  <Step title="Save">
    Click **Save** to start the sync job.
  </Step>
</Steps>

## Related

* [Fix lyrics sync](/en/desktop/fix-lyrics-sync) if timing still needs fine-tuning
* [Edit the lyrics](/en/desktop/edit-lyrics) to correct text before re-syncing