[Integrations](https://spiral.ai/integrations) › [Claude](https://spiral.ai/integrations/claude-desktop) › [Zonos2](https://spiral.ai/integrations/claude-desktop/zonos2)

# How to Use Zonos2 in Claude

Zonos2 turns scripts into multilingual speech with natural voice cloning, for narration, spoken updates, and dialogue previews. Its accuracy mode favors a closer voice match; switching it off allows more expression.

|  |  |
| --- | --- |
| Model | Zonos2 |
| Available through | fal |
| Category | Audio |
| Price | Up to $0.013 per run at the default inputs for 60 seconds of input. |
| Returns | audio |
| Runs | In the background |
| Setup | About 2 minutes, no API key |

## Prompts to try with Zonos2 in Claude

- Write a 70-word welcome message for a neighborhood gardening workshop, then read it with Zonos2 using my supplied voice reference. Prioritize a close voice match.
- Prepare English and Japanese versions of a two-sentence museum welcome. Generate each separately with Zonos2 using my supplied voice reference, and show me both scripts.
- Use Zonos2 to read this exact line twice with my supplied voice reference: 'You came back. I thought I had missed my chance.' Prioritize voice accuracy for the first reading and expression for the second.
- Create a short English explanation of the Mandarin greeting 你好, including the greeting in Mandarin, and generate it with Zonos2. Keep the script below 50 words and show the text before the recording.
- Shorten my announcement to 80 words, spell out dates and numbers as they should be spoken, and generate it with Zonos2 using my supplied voice reference.

## How to set up Zonos2 in Claude

Pick the way that suits you. Each takes about two minutes.

### Use the Claude connector

1. **Add Spiral to Claude**: Click Add Spiral in Claude above. Claude opens Add custom connector with Spiral and its URL filled in; check them and click Continue. Without the link: in Claude open Customize › Connectors › Add › Add custom connector and paste this URL:

```text
https://api.spiral.ai/mcp
```

2. **Keep the detected sign-in and click Add**: Claude finds Spiral's sign-in by itself: leave Sign in now and Use Claude's published identity selected.

3. **Click Connect**: The connector page opens. Connect takes you to Spiral in your browser.

### Use the Spiral skill

1. **Send Claude this message**: Your agent reads Spiral's skill and sets itself up.

```
set up https://spiral.ai/SKILL.md
```

2. **Sign in or paste a key**: Approve Spiral in the browser, or give the agent a key from the console.

### Use the Spiral CLI

1. **Install and sign in**: Installs the Spiral CLI, signs you in and adds its skill to your coding agents.

```bash
npx -y spiral-cli@latest init
```

### Call the Spiral API

1. **Create an API key**: In the console, under API keys.

2. **Run Zonos2**: One request; the result comes back by polling or a webhook.

```bash
curl https://api.spiral.ai/v1/runs \
  -H "Authorization: Bearer $SPIRAL_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{
    "provider": "fal",
    "model": "fal-ai/zonos2",
    "input": {
      "reference_audio_url": "https://storage.googleapis.com/falserverless/model_tests/zonos/demo_voice_zonos.wav"
    }
  }'
```

Then paste one of the prompts above: Claude finds Zonos2 on Spiral, shows the price and runs it.

Setup guide in the docs: https://spiral.ai/docs/quickstarts/claude-desktop

## Zonos2 tools and prices

- [Zonos2 Text to Speech](https://spiral.ai/tools/fal/fal-ai/zonos2) (fal): $0.013 per minute.

## Settings you can ask for

| Setting | Allowed values | Default | Description |
| --- | --- | --- | --- |
| text |  | "Fal" is the fastest solution for your audio generation. | Text to synthesize in the cloned voice. |
| language |  | en_us | Text-normalization language code. Supported: en_us, en_gb, fr_fr, de, es, it, pt_br, ja, cmn, ko. Unsupported codes skip normalization. |
| seed | 0 to 2147483647 | — | Seed for reproducibility. Random when omitted. |
| accurate_mode |  | true | True = closer voice match; False = more expressive. |
| clean_speaker_background |  | false | Mark the reference audio as having a clean background. |
| temperature | 0 to 2 | 1.15 | Sampling temperature. |

## What can you do with Zonos2 in Claude?

Turn an approved script into narration or a multilingual voice recording with Zonos2. A chat agent can shorten the text for a one-minute passage and help compare voice accuracy with expression.

### Product walkthrough narration

Turn a short walkthrough script into speech with Zonos2. Generate each section separately so you can replace a line when the product changes.

### Multilingual announcements

Produce separate English, Mandarin Chinese, and Japanese readings of an approved announcement with Zonos2. Review each language before distributing the recordings.

### Dialogue auditions

Use Zonos2 to audition short fictional dialogue with accuracy mode on and off. Compare voice resemblance and expression before choosing a reading.

### Audio learning cards

Create spoken examples for vocabulary cards with Zonos2. Keep each card to a short sentence and check pronunciation before adding it to a lesson.

### Daily spoken briefings

Convert a concise daily summary into a Zonos2 recording under one minute. Prepare the summary first, then synthesize only the approved spoken text.

### Podcast intros and corrections

Generate a short introduction or replacement passage with Zonos2 using the host's supplied voice reference. Listen for differences in recording texture before editing it into an episode.

### Bilingual explanations

Use Zonos2 to read a short explanation that switches between languages, such as an English lesson containing Mandarin examples. Check the words around each language change.

## Overview

- [fal](https://spiral.ai/providers/fal)
- [All Claude guides](https://spiral.ai/integrations/claude-desktop)

## Zonos2 in other agents

- [Amp](https://spiral.ai/integrations/amp/zonos2)
- [ChatGPT](https://spiral.ai/integrations/chatgpt/zonos2)
- [Claude Code](https://spiral.ai/integrations/claude-code/zonos2)
- [Cline](https://spiral.ai/integrations/cline/zonos2)
- [Codex](https://spiral.ai/integrations/codex/zonos2)
- [Cursor](https://spiral.ai/integrations/cursor/zonos2)
- [Devin](https://spiral.ai/integrations/devin/zonos2)
- [dots](https://spiral.ai/integrations/dots/zonos2)
- [Factory Droid](https://spiral.ai/integrations/factory-droid/zonos2)
- [Gemini](https://spiral.ai/integrations/gemini/zonos2)
- [Gemini CLI](https://spiral.ai/integrations/gemini-cli/zonos2)
- [Google Antigravity](https://spiral.ai/integrations/antigravity/zonos2)
- [goose](https://spiral.ai/integrations/goose/zonos2)
- [Grok](https://spiral.ai/integrations/grok/zonos2)
- [Grok Build](https://spiral.ai/integrations/grok-build/zonos2)
- [Hermes](https://spiral.ai/integrations/hermes/zonos2)
- [Instinct](https://spiral.ai/integrations/instinct/zonos2)
- [Manus](https://spiral.ai/integrations/manus/zonos2)
- [Muse](https://spiral.ai/integrations/muse/zonos2)
- [Muse Code](https://spiral.ai/integrations/muse-code/zonos2)
- [n8n](https://spiral.ai/integrations/n8n/zonos2)
- [Notion](https://spiral.ai/integrations/notion/zonos2)
- [OpenClaw](https://spiral.ai/integrations/openclaw/zonos2)
- [opencode](https://spiral.ai/integrations/opencode/zonos2)
- [Perplexity](https://spiral.ai/integrations/perplexity/zonos2)
- [Raycast](https://spiral.ai/integrations/raycast/zonos2)
- [VS Code](https://spiral.ai/integrations/vscode/zonos2)
- [Windsurf](https://spiral.ai/integrations/windsurf/zonos2)
- [Zapier](https://spiral.ai/integrations/zapier/zonos2)

## More Claude guides

- [Whisper Large](https://spiral.ai/integrations/claude-desktop/whisper-large)
- [xAI Text to Speech](https://spiral.ai/integrations/claude-desktop/xai-text-to-speech)
- [VEED Clean Audio](https://spiral.ai/integrations/claude-desktop/veed-clean-audio)
- [Bria Fibo Lite structured prompt](https://spiral.ai/integrations/claude-desktop/bria-fibo-lite-structured-prompt)
- [Universal-3.5 Pro](https://spiral.ai/integrations/claude-desktop/universal-3-5-pro)
- [Bria Fibo structured edit instructions](https://spiral.ai/integrations/claude-desktop/bria-fibo-structured-edit-instructions)

## Questions about Zonos2 in Claude

**Can Claude use Zonos2?**

Yes. Add Spiral to Claude as an MCP server, and Claude can find and run Zonos2 like any built-in tool.

**How much does Zonos2 cost in Claude?**

Up to $0.013 per run at the default inputs for 60 seconds of input. You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your Claude plan is separate.

**Do I need a fal account to use Zonos2 in Claude?**

No. Spiral runs Zonos2 on its own fal account. You sign in to Spiral once and never paste a key.

**How long does it take to set up Zonos2 in Claude?**

About two minutes: add Spiral to Claude, then keep the detected sign-in and click Add, then click Connect.

**Does Claude wait while Zonos2 runs?**

Zonos2 runs in the background. Claude starts the run, checks on it and gives you the audio when it's ready.

**What is Zonos2 best used for?**

Zonos2 suits spoken content where the sound of a particular voice matters: short narration, dialogue previews, and multilingual announcements. It can clone a voice from a short sample.

**How long can a Zonos2 clip be?**

Zyphra documents generation up to one minute. For longer narration, split the script into separate passages and review the joins.

**Which languages should I try first?**

English, Mandarin Chinese, and Japanese occupy Zyphra's highest support tier. French, Spanish, German, and several other languages are supported in a lower tier, so audition your actual script before committing to a batch.

**Should I prioritize voice accuracy or expression?**

Keep accuracy mode on when matching the reference voice matters most. Try it off for a more expressive reading, then compare both recordings with the same script.

**Will Zonos2 clean up a noisy voice reference?**

Do not assume it will remove every imperfection. Zyphra emphasizes vocal fidelity even with noisy or distorted references; choose a clean recording when clean narration is the goal.

