# AI use cases for Yovoice (leemysw/yovoice)

Repo: [leemysw/yovoice](https://github.com/leemysw/yovoice) | Use cases: 5 | Page: https://dedupedai.com/use-cases/leemysw-yovoice/

yovoice is an open-source, local-first studio for macOS and Windows that generates speech, cloned voices, songs with lyrics, and multi-track MIDI scores on your own machine.

This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL.

- Turn a text file into narration with an AI agent: https://dedupedai.com/use-cases/leemysw-yovoice/narrate-text-file-with-agent/
- Dub a script or subtitle file line by line: https://dedupedai.com/use-cases/leemysw-yovoice/dub-story-lines-by-character/
- Generate a sung track from your own lyrics offline: https://dedupedai.com/use-cases/leemysw-yovoice/write-song-from-lyrics-locally/
- Arrange a multi-part score and export MIDI and stems: https://dedupedai.com/use-cases/leemysw-yovoice/arrange-midi-score-with-agent/
- Expose local audio generation as an API and MCP server: https://dedupedai.com/use-cases/leemysw-yovoice/serve-local-audio-api-for-coding-agent/

## How to turn a text file into narration with an AI agent with Yovoice

Repo: [leemysw/yovoice](https://github.com/leemysw/yovoice) | Difficulty: Intermediate | Time: 1 hour | Basis: From the README

An agent reads a script file and saves a narrated WAV using your reference voice, with no desktop app open.

**Who it is for:** A video or course creator who keeps scripts in plain text files and wants an agent to produce audio files from them.

**The problem:** Recording narration by hand is slow, and cloud text-to-speech bills per character or per minute.

**The outcome:** A WAV file of the script in the chosen delivery style, generated locally by the Yovoice CLI.

### Steps

1. Give your agent the Yovoice Skill link (https://github.com/leemysw/yovoice/tree/main/skills/yovoice) and ask it to install the Skill and set up the local CLI, engine, and model.
2. Place narration.txt and a reference voice file voice.wav in the working folder.
3. Ask the agent to read narration.txt using voice.wav as the reference voice, with a calm delivery, and save it as narration.wav.
4. Ask the agent to report which model and engine it used, so you can repeat the run. (inferred, check the README)
5. Play narration.wav and check that the pacing and pronunciation match the script.

### Prompt to paste into your agent

```text
Using the leemysw/yovoice repo and its Skill at skills/yovoice, read <script.txt> with <reference.wav> as the reference voice in a calm delivery and save the result as <output.wav>. Report the model used and confirm <output.wav> exists and plays back at the expected script length.
```

### Caveat

The README does not give the standalone CLI install command, so the agent must follow the Skill instructions and the docs/cli.md guide.

Risk, voice cloning: Only use a reference voice recording from someone who has given you permission to clone it.

Dual-use notice: only use this on software, sites and systems you own or are authorized to test.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/leemysw-yovoice/narrate-text-file-with-agent/

## How to dub a script or subtitle file line by line with Yovoice

Repo: [leemysw/yovoice](https://github.com/leemysw/yovoice) | Difficulty: Intermediate | Time: 1 hour | Basis: From the README

Import text or subtitles, assign a voice to each character, and generate every line in the desktop app.

**Who it is for:** A short-film maker, game prototyper, or language teacher with a dialogue script and several speaking roles.

**The problem:** Assigning and rendering each speaking part by hand across a whole script is tedious to do in separate tools.

**The outcome:** A set of audio lines, one per script entry, each sung or spoken by its assigned character voice.

### Steps

1. Download the .dmg (macOS) or -setup.exe (Windows) from the Releases tab or the official website and install it.
2. Open Settings and download a model, for example Kokoro-82M for fast narration.
3. Create a project and choose Story dubbing.
4. Import your text or subtitle file, then assign a character to each line.
5. Select Generate, then preview and compare takes before exporting.

### Prompt to paste into your agent

```text
Using the leemysw/yovoice repo, help me set up a Story dubbing project from <script file> with one voice per character in <character list>. Expected output is one generated audio take per line; verify that every line in the script has a generated take before export.
```

### Caveat

Built-in voices come from Kokoro-82M and Qwen3-TTS CustomVoice, so a custom cloned voice needs a cloning model such as VoxCPM2 or IndexTTS 2.0.

Dual-use notice: only use this on software, sites and systems you own or are authorized to test.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/leemysw-yovoice/dub-story-lines-by-character/

## How to generate a sung track from your own lyrics offline with Yovoice

Repo: [leemysw/yovoice](https://github.com/leemysw/yovoice) | Difficulty: Beginner | Time: 1 hour | Basis: From the README

Write a style and lyrics with section markers, then generate a song with ACE-Step 1.5 in the local studio.

**Who it is for:** A songwriter or hobby musician who has lyrics but no recording setup and does not want to pay for a music subscription.

**The problem:** Turning finished lyrics into a demo usually needs a vocalist, a studio, or a paid cloud music service.

**The outcome:** A generated song with vocals and an instrumental backing, with lyrics synced to the vocals on the playback page.

### Steps

1. Install the Yovoice package for your platform from the Releases tab or the official website.
2. Open Settings and download ACE-Step 1.5 Turbo.
3. Create a Music project, then write a style description and your lyrics with section markers.
4. Select Generate and listen on the playback page, where lyrics are synced to the vocals.
5. Compare takes side by side and export the one you like.

### Prompt to paste into your agent

```text
Using the leemysw/yovoice repo, write a Music project with the style <style description> and the lyrics in <lyrics file>, using ACE-Step 1.5. Expected output is a generated song with synced lyrics; verify that all section markers from the lyrics file appear in the playback view.
```

### Caveat

YuE2 weights are restricted to non-commercial use, so check the model license before using YuE2 output commercially; ACE-Step 1.5 is the route named for songs here.

Dual-use notice: only use this on software, sites and systems you own or are authorized to test.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/leemysw-yovoice/write-song-from-lyrics-locally/

## How to arrange a multi-part score and export MIDI and stems with Yovoice

Repo: [leemysw/yovoice](https://github.com/leemysw/yovoice) | Difficulty: Intermediate | Time: 1 hour | Basis: From the README

Write parts by bar and beat, import MIDI or let an agent compose, render with a General MIDI sound font, and export.

**Who it is for:** A composer or arranger who wants to sketch a multi-part piece and hand the MIDI and audio stems to a DAW.

**The problem:** Getting a rough arrangement rendered and exported as MIDI and stems often means juggling several separate tools.

**The outcome:** A rendered score with an exported MIDI file and audio stems for each part.

### Steps

1. Install Yovoice and open a Score project.
2. Write parts by bar and beat, or import an existing MIDI file.
3. Ask the built-in assistant or your agent to draft a missing part, if you want an agent to compose. (inferred, check the README)
4. Render the score using the General MIDI sound font built into the app.
5. Export the MIDI file and the stems.

### Prompt to paste into your agent

```text
Using the leemysw/yovoice repo, create a Score project with the parts described in <arrangement brief> and export the MIDI and stems to <output folder>. Expected output is one MIDI file and one stem per part; verify that each part's bar count matches the brief.
```

### Caveat

The README says the Score feature can import MIDI or let an Agent compose, but it does not describe the exact arrangement controls, so follow docs/score.md for those details.

Dual-use notice: only use this on software, sites and systems you own or are authorized to test.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/leemysw-yovoice/arrange-midi-score-with-agent/

## How to expose local audio generation as an API and MCP server with Yovoice

Repo: [leemysw/yovoice](https://github.com/leemysw/yovoice) | Difficulty: Advanced | Time: 2 hours | Basis: From the README

Run Yovoice serve so Cursor or Claude Code can call the authenticated HTTP API and MCP inference service.

**Who it is for:** A developer who wants a coding agent in Cursor or Claude Code to generate speech or music on a machine with a capable GPU.

**The problem:** Coding agents cannot generate audio locally without a service they can reach over the network.

**The outcome:** An authenticated HTTP API and MCP endpoint that the agent client can call for inference.

### Steps

1. Install Yovoice on the machine that has the models and GPU, then download the models you need.
2. Run `yovoice serve` on that machine.
3. Follow the API / MCP deployment guide at docs/api.md to configure authentication and connect the client.
4. Add the MCP server to Cursor or Claude Code using the client example in docs/api.md.
5. Ask the agent for a short test clip and confirm it returns audio.

### Prompt to paste into your agent

```text
Using the leemysw/yovoice repo, run `yovoice serve` on <host machine> and connect Claude Code or Cursor to its MCP endpoint following docs/api.md. Expected output is a short test audio clip returned by the agent; verify that unauthenticated requests are rejected.
```

### Caveat

The README gives the serve command but not its authentication setup, so the exact configuration must be taken from docs/api.md.

Risk, credentials: The HTTP API is authenticated, so keep its credentials out of shared repos and restrict who on the network can reach the host.

Dual-use notice: only use this on software, sites and systems you own or are authorized to test.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/leemysw-yovoice/serve-local-audio-api-for-coding-agent/