ElevenLabs Multilingual V2 & Turbo V2.5

Create natural multilingual speech with ElevenLabs through the AI Video API asynchronous audio endpoint. Both models support built-in voices, adjustable voice similarity and stability, and inline pauses.

Models and pricing

Billing uses the number of Unicode code points in input.text. Every request has a 1,000-character minimum charge, while the input limit is 10,000 characters.

Model	Credits per character	Minimum charge
`elevenlabs/multilingual-v2`	0.03	30 credits
`elevenlabs/turbo-v2.5`	0.015	15 credits

Emoji and other supplementary Unicode characters each count as one character. Use POST /v1/estimate to calculate the same charge that will be pre-authorized when the task is created.

Create a task

curl -X POST https://api.aivideoapi.ai/v1/audio/generations \
  -H "Authorization: Bearer sk-your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "elevenlabs/multilingual-v2",
    "callback_url": "https://your-server.com/webhook",
    "input": {
      "text": "Welcome to ElevenLabs.<#0.8#> Let us begin.",
      "voice_id": "Alice",
      "similarity": 1,
      "stability": 0.5,
      "use_speaker_boost": true
    }
  }'

Request body

Field	Type	Required	Description
`model`	string	Yes	`elevenlabs/multilingual-v2` or `elevenlabs/turbo-v2.5`
`input`	object	Yes	Text-to-speech parameters described below
`callback_url`	string	No	HTTPS URL that receives the final task payload

Input

Field	Type	Required	Default	Description
`text`	string	Yes	—	Text to synthesize, from 1 to 10,000 Unicode characters. Insert `<#x#>` between words to pause for `x` seconds (`0.01`–`99.99`)
`voice_id`	string	Yes	—	A built-in voice name below, or another non-empty voice identifier supported by your request
`similarity`	number	No	`1`	Target-speaker similarity from `0` to `1`
`stability`	number	No	`0.5`	Voice stability from `0` to `1`
`use_speaker_boost`	boolean	No	`true`	Improves clarity for English numbers, dates, and units

Built-in voices

The Playground defaults to Alice. Preview any locally hosted voice below, then use its name as voice_id.

voice_id	Description	Preview
`Alice`	Clear, engaging British woman for e-learning
`Aria`	Calm, lightly raspy middle-aged woman
`Bill`	Friendly and comforting storyteller
`Brian`	Resonant, comforting narration and advertising voice
`Callum`	Gravelly voice with an unsettling edge
`Charlie`	Confident and energetic young Australian man
`Charlotte`	Sensual, raspy character voice
`Chris`	Natural, down-to-earth all-purpose voice
`Daniel`	Strong professional broadcast and news voice
`Eric`	Smooth tenor suited to agentic use cases
`George`	Warm and captivating resonance
`Jessica`	Playful young American woman for trendy content
`Laura`	Sunny, enthusiastic young woman with a quirky attitude
`Liam`	Energetic, warm young adult for reels and shorts
`Lily`	Warm, clear British woman for news and narration
`Matilda`	Professional woman with a pleasing alto pitch
`River`	Relaxed, neutral narration and conversation voice
`Roger`	Easygoing voice for casual conversations
`Sarah`	Warm, confident and reassuring professional woman
`Will`	Conversational and laid-back voice

Create response

{
  "code": 200,
  "msg": "success",
  "data": {
    "taskId": "a1b2c3d4-e5f6-7890-abcd-ef1234567890"
  }
}

Estimate credits

This request only calculates credits and does not create a task:

curl -X POST https://api.aivideoapi.ai/v1/estimate \
  -H "Authorization: Bearer sk-your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "elevenlabs/turbo-v2.5",
    "input": {
      "text": "Hello 👋",
      "voice_id": "Alice"
    }
  }'

Query a task

curl https://api.aivideoapi.ai/v1/tasks/{taskId} \
  -H "Authorization: Bearer sk-your-api-key"

Status transitions are pending → processing → completed or failed.

Completed

{
  "id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
  "status": "completed",
  "model": "elevenlabs/multilingual-v2",
  "created_at": 1786415700,
  "completed_at": 1786415704,
  "output": {
    "urls": [
      "https://file.aivideoapi.ai/audio/example.mp3"
    ],
    "metadata": {
      "model": "elevenlabs/multilingual-v2"
    }
  }
}

Audio URLs are valid for 24 hours.

Failed

{
  "id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
  "status": "failed",
  "model": "elevenlabs/multilingual-v2",
  "created_at": 1786415700,
  "error": {
    "code": "upstream_error",
    "message": "Audio generation failed"
  }
}

A confirmed failed task is refunded automatically and at most once. If task submission is temporarily uncertain, the task remains pending while AI Video API verifies its status. These fast TTS tasks cannot be cancelled; cancellation requests return HTTP 409.

Callback

Set callback_url in the create request. AI Video API sends a POST request when the task completes or fails. The callback body matches the task query payload and uses the same sanitized public error message. If delivery fails, query GET /v1/tasks/{taskId} as the fallback.