# ElevenLabs v4 Turbo Text-to-Speech — Atlas Cloud API

> ElevenLabs' low-latency voice model: the expressiveness, inline delivery tags and multi-speaker dialogue of ElevenLabs v4 at half the price.

This is the machine-readable API reference for **ElevenLabs v4 Turbo Text-to-Speech** on Atlas Cloud,
a unified API platform for 400+ AI models across text, image, video, audio and 3D.

- **Model ID**: `elevenlabs/v4-turbo/text-to-speech`
- **Built by**: ELEVENLABS
- **Modality**: Audio
- **Model page**: https://www.atlascloud.ai/models/elevenlabs/v4-turbo/text-to-speech
- **API key**: https://www.atlascloud.ai/console/api-keys
- **Docs**: https://www.atlascloud.ai/docs

## Pricing on Atlas Cloud

- $0.04 per 1K characters
- Pay-as-you-go. No minimum spend, no subscription required.

> **These are the authoritative Atlas Cloud rates for this model.** Any price that
> appears in the vendor description further down refers to a different platform or
> a different model variant and does not apply here.

## Use this model from an AI agent

Atlas Cloud ships three first-party integration surfaces. All three authenticate
with the same API key via the `ATLASCLOUD_API_KEY` environment variable.

### MCP server

The official MCP server (`atlascloud-mcp`) exposes this model to any
MCP-compatible host — Claude Code, OpenAI Codex, Cursor, Gemini CLI, Goose,
Claude Desktop. One-line install:

```bash
# Claude Code
claude mcp add atlascloud -- npx -y atlascloud-mcp

# OpenAI Codex CLI
codex mcp add atlascloud -- npx -y atlascloud-mcp

# Gemini CLI
gemini mcp add atlascloud -- npx -y atlascloud-mcp

export ATLASCLOUD_API_KEY="your-api-key"
```

Then ask in plain English; the agent calls `atlas_generate_audio` with `model: "elevenlabs/v4-turbo/text-to-speech"`.
The server fetches each model's schema and validates parameters before submitting,
so invalid requests fail fast without spending credits.

MCP docs: https://www.atlascloud.ai/docs/mcp-server

### Agent Skills

`atlas-cloud-skills` is a portable skill package (API reference, code templates in
Python / Node.js / cURL, model IDs with pricing) for Claude Code, Cursor, Codex and
12+ other agents:

```bash
npx skills add AtlasCloudAI/atlas-cloud-skills
export ATLASCLOUD_API_KEY="your-api-key"
```

Skills docs: https://www.atlascloud.ai/docs/skills

### CLI

The `atlas` binary runs Atlas Cloud from a terminal or CI script. Async media jobs
are polled and downloaded automatically (use `--no-download` when a script only
needs the output URLs):

```bash
# Install (Homebrew, npm, or shell installer)
brew install AtlasCloudAI/tap/atlascloud
# npm install -g atlascloud-cli
# curl -fsSL https://raw.githubusercontent.com/AtlasCloudAI/cli/main/install.sh | sh

atlas auth login
atlas models get elevenlabs/v4-turbo/text-to-speech --json
```

CLI docs: https://www.atlascloud.ai/docs/cli

## HTTP API reference

- **Submit endpoint (POST)**: `https://api.atlascloud.ai/api/v1/model/generateAudio` — start an async generation; returns a `prediction_id`
- **Poll endpoint (GET)**: `https://api.atlascloud.ai/api/v1/model/prediction/{prediction_id}` — poll this until the prediction finishes
- **Model ID**: `elevenlabs/v4-turbo/text-to-speech`


## API Information

This model can be used via our HTTP API or more conveniently via our client libraries.
See the input and output schema below, as well as the usage examples.


### Input Schema

The API accepts the following input parameters:

- **`model`** (`string`, _required_):
  Model name.
  - Default: `"elevenlabs/v4-turbo/text-to-speech"`

- **`text`** (`string`, _optional_):
  Text to synthesize into speech audio. Supports inline delivery tags such as [cheerfully], [whispers] and [sighs]. Ignored when `inputs` is given. Hard limit is 10,000 characters. ElevenLabs recommends staying at or below 2,000 characters for reliable generation, but the real constraint is that the upstream connection is cut after about 60 seconds of synthesis; Turbo runs ~4x faster than v4 and in testing handles up to ~5,000 characters, so keep a request below that to avoid a mid-stream failure.
  - Default: `"Welcome to Atlas Cloud - One API for All Media AI."`

- **`voice`** (`string`, _optional_):
  Voice for the speech. Pick one of the preset names below, or pass a raw ElevenLabs voice_id (including your own cloned voices), which is forwarded as-is. When `inputs` is given, this is the fallback voice for turns that name none.
  - Default: `"Roger"`
  - Options: "Adam", "Alice", "Bella", "Bill", "Brian", "Callum", "Charlie", "Chris", "Daniel", "Eric", "George", "Harry", "Jessica", "Laura", "Liam", "Lily", "Matilda", "River", "Roger", "Sarah", "Will"

- **`output_format`** (`string`, _optional_):
  Output audio format and bitrate. Higher MP3 bitrates and 44.1 kHz PCM/WAV require a higher ElevenLabs plan.
  - Default: `"mp3_44100_128"`
  - Options: "mp3_44100_128", "mp3_44100_64", "mp3_22050_32", "opus_48000_128", "pcm_24000", "pcm_16000", "wav_44100", "ulaw_8000", "alaw_8000"

- **`stability`** (`number`, _optional_):
  How steady the delivery is. Lower is more expressive and more variable.
  - Default: `0.5`
  - Min: 0
  - Max: 1

- **`similarity_boost`** (`number`, _optional_):
  How closely the output tracks the original voice.
  - Default: `0.75`
  - Min: 0
  - Max: 1

- **`clone_strength`** (`number`, _optional_):
  Strength of the voice clone. Applies only to professional voice clones.
  - Min: 0
  - Max: 1

- **`language_code`** (`string`, _optional_):
  Language code (ISO 639-1) used to enforce a language for the model.
  - Default: `null`

- **`apply_text_normalization`** (`string`, _optional_):
  Whether numbers, dates and abbreviations are spelled out before synthesis.
  - Default: `"auto"`
  - Options: "auto", "on", "off"

- **`seed`** (`integer`, _optional_):
  Seed for best-effort reproducible output. Leave empty for a random seed.
  - Min: 0
  - Max: 4294967295

- **`inputs`** (`array[object]`, _optional_):
  Multi-speaker dialogue, one object per turn. Overrides `text` and `voice`. At most 10 distinct voices per request, and 10,000 characters across all turns (hard limit). ElevenLabs recommends 2,000 characters combined for reliable generation; the real constraint is the ~60-second upstream cutoff, so in practice keep the combined text at or below ~5,000 characters.
  - Item properties:
    - **`text`** (`string`, _required_):
      What this turn says. Supports the same inline delivery tags as `text`.

    - **`voice`** (`string`, _optional_):
      Preset voice name or raw voice_id for this turn. Falls back to the top-level `voice`.
      - Options: "Adam", "Alice", "Bella", "Bill", "Brian", "Callum", "Charlie", "Chris", "Daniel", "Eric", "George", "Harry", "Jessica", "Laura", "Liam", "Lily", "Matilda", "River", "Roger", "Sarah", "Will"




**Required Parameters Example**:

```json
{
  "model": "elevenlabs/v4-turbo/text-to-speech"
}
```


**Full Example**:

```json
{
  "model": "elevenlabs/v4-turbo/text-to-speech",
  "text": "Welcome to Atlas Cloud - One API for All Media AI.",
  "voice": "Roger",
  "output_format": "mp3_44100_128",
  "stability": 0.5,
  "similarity_boost": 0.75,
  "clone_strength": 0,
  "language_code": null,
  "apply_text_normalization": "auto",
  "seed": 0,
  "inputs": [
    {
      "text": "",
      "voice": "Adam"
    }
  ]
}
```


### Output Schema

The API returns the following output format:


- **`id`** (`string`, _optional_):
  Unique identifier for the prediction.

- **`model`** (`string`, _optional_):
  Model ID used for the prediction.

- **`status`** (`string`, _optional_):
  Status of the task: created, processing, completed, or failed.

- **`outputs`** (`array[string]`, _optional_):
  Array of URLs to the generated audio.

- **`created_at`** (`string`, _optional_):
  ISO timestamp of when the request was created.

- **`urls`** (`object`, _optional_):
  Object containing related API endpoints.



**Example Response**:

```json
{
  "id": "",
  "model": "",
  "status": "",
  "outputs": [
    ""
  ],
  "created_at": "",
  "urls": {}
}
```


## Usage Examples

### cURL

```bash
# Step 1: Start generation (async)
curl -X POST "https://api.atlascloud.ai/api/v1/model/generateAudio" \
  -H "Authorization: Bearer $ATLASCLOUD_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "elevenlabs/v4-turbo/text-to-speech",
  "text": "Welcome to Atlas Cloud - One API for All Media AI.",
  "voice": "Roger",
  "output_format": "mp3_44100_128",
  "stability": 0.5,
  "similarity_boost": 0.75,
  "clone_strength": 0,
  "language_code": null,
  "apply_text_normalization": "auto",
  "seed": 0,
  "inputs": [
    {
      "text": "",
      "voice": "Adam"
    }
  ]
}'

# Response will contain: {"code": 200, "data": {"id": "prediction_id", "status": "processing"}}

# Step 2: Poll for result (replace {prediction_id} with the id returned above)
curl -X GET "https://api.atlascloud.ai/api/v1/model/prediction/{prediction_id}" \
  -H "Authorization: Bearer $ATLASCLOUD_API_KEY"

# Keep polling until status is "completed", "succeeded" or "failed"
# When completed, outputs will contain the generated content URL(s)
```

## Additional Resources

### Documentation

- [Model Playground](https://www.atlascloud.ai/models/elevenlabs/v4-turbo/text-to-speech)

## About this model

_Vendor-supplied description. Any pricing or endpoint mentioned below refers to_
_other platforms — use the Atlas Cloud values above._

### ElevenLabs v4 Turbo Text-to-Speech

**ElevenLabs v4 Turbo** is the low-latency member of the ElevenLabs v4 family. It keeps the expressiveness, inline delivery tags and multi-speaker dialogue of v4, runs several times faster, and costs roughly half as much — a fit for real-time and high-volume synthesis.

#### Why Choose This?

- **Low latency.** Built for responsive, real-time speech — voice agents, live assistants and interactive apps.
- **v4 expressiveness.** Same inline delivery tags (`[cheerfully]`, `[whispers]`, `[sighs]`, `[laughs]`) and natural intonation as v4.
- **Multi-speaker dialogue.** Pass `inputs` with one object per turn to stage a back-and-forth between different voices in one request.
- **Preset or cloned voices.** Pick a preset name, or pass any ElevenLabs `voice_id` (including your own cloned voices) and it is forwarded as-is.
- **Lower cost.** Roughly half the price of v4 for the same request shape. Billed per 1,000 characters of input text.

#### Parameters

| Parameter | Required | Description |
| --- | --- | --- |
| model | Yes | `elevenlabs/v4-turbo/text-to-speech` |
| text | Yes* | Text to synthesize, up to 10,000 characters. Supports inline delivery tags. Ignored when `inputs` is given. |
| voice | No | Preset voice name or a raw ElevenLabs `voice_id`. Default `Roger`. When `inputs` is given, this is the fallback voice for turns that name none. |
| inputs | No | Multi-speaker dialogue: an array with one `{ "text": ..., "voice": ... }` object per turn. Overrides `text`/`voice`. Up to 10 distinct voices; 10,000 characters combined. |
| output_format | No | `mp3_44100_128` (default), `mp3_44100_64`, `mp3_22050_32`, `opus_48000_128`, `pcm_24000`, `pcm_16000`, `wav_44100`, `ulaw_8000`, `alaw_8000`. Higher MP3 bitrates and 44.1 kHz PCM/WAV require a higher ElevenLabs plan. |
| stability | No | 0–1. How steady the delivery is; lower is more expressive and variable. Default `0.5`. |
| similarity_boost | No | 0–1. How closely the output tracks

_(Description truncated. Full text on the model page.)_

---

Atlas Cloud — one API for 400+ AI models. Model page: https://www.atlascloud.ai/models/elevenlabs/v4-turbo/text-to-speech · Docs: https://www.atlascloud.ai/docs