# Generate AI Voiceovers with inference.sh

Teams need fast narration without building audio tooling. This skill guides Claude, Codex, and Claude Code through inference.sh TTS workflows.

## Install

```bash
npx skillstore add inference-sh/ai-voice-cloning
```

## Metadata

- Status: approved
- Slug: inference-sh-ai-voice-cloning
- Skillstore revision: r1
- Version status: missing
- Tree hash: 877e5fa3ac58542c0318f7922afac2ae33c3de3ad26ad738b3cd50e6b0ca5739
- Author: inference-sh
- GitHub username: inference-sh
- License: MIT
- Repository: https://github.com/inference-sh/skills/tree/main/skills/ai-voice-cloning/
- Ref: a06681402992ceae98ba04d54cfd4ab004862696
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: blocked
- Manual install advisory: allowed\_with\_warning
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: external\_commands, network
- Quality score: 38
- Quality tier: warning
- Public page: https://skillstore.pages.dev/skills/inference-sh-ai-voice-cloning
- Manifest: https://skillstore.pages.dev/api/skills/inference-sh-ai-voice-cloning/manifest

## Capabilities

- Lists supported inference.sh TTS models and common voice choices.
- Provides infsh CLI examples for narration, dialogue, podcast, and video voiceover tasks.
- Shows chunking and media-merger workflows for long-form audio.
- Explains speed and punctuation controls for pacing.
- Connects generated speech to avatar and video workflows.

## Use Cases

- Create video narration: Plan voiceover text, voice choice, pacing, and media merge steps for short videos.
- Produce learning audio: Convert course scripts into clear narration settings for lessons and accessibility content.
- Prototype podcast segments: Draft multi-voice dialogue workflows and merge steps for sample podcast episodes.

## Prompt Templates

### Generate a short voiceover

```
Use this skill to create a concise inference.sh voiceover plan for this script: [script]. Recommend one voice and speed.
```

### Match a voice style

```
Compare suitable voices for [audience] and [tone]. Explain the best choice and pacing settings for the script.
```

### Plan long-form narration

```
Break this long script into narration chunks, keep the voice consistent, and outline merge steps for final audio.
```

### Design a multi-speaker workflow

```
Create a multi-speaker production plan for [project]. Include voice choices, chunking, merge order, and consent checks.
```

## Limitations

- Requires the inference.sh CLI, account login, and network access.
- Provides documentation examples; it does not generate audio by itself.
- Does not include consent checks for cloned voices or avatar likenesses.
- Uses hosted models, so sensitive text and media need review before use.

## Best Practices

- Confirm rights and consent before using any real voice, likeness, or portrait.
- Test several voices and speeds before producing the final audio.
- Review scripts for private, regulated, or third-party content before sending them to hosted models.

## Anti Patterns

- Do not imitate a real person without permission and clear disclosure.
- Do not send confidential scripts or private media to hosted models without approval.
- Do not run remote installer commands without verifying the source and integrity.

## Security Audit

- Audited at: 2026-07-06T15:59:40.981\+00:00
- Summary: The audit confirms the quick start recommends a remote installer piped directly to sh, which is a critical supply-chain risk. Most external command alerts are Markdown fences, inline model IDs, or documented infsh examples, not executable skill code. The skill also has dual-use risk because it documents synthetic voice and talking-head workflows without consent or disclosure guidance.

## Stats

- Views: 167
- Downloads: 10
- Favorites: 0
- Popularity score: 0