Gemini 3.1 Flash TTS

Model ID

Use with the Workers AI binding, the AI Gateway REST API, or the Unified API.

Calling the provider directly? Use the provider's model name instead, including with provider-native endpoints or the provider's own SDK. For example, Anthropic's name for anthropic/claude-opus-4.8 is claude-opus-4-8.

Third-party
Zero data retention

Model Info
Zero data retention	Yes
Pricing	Input audio (per 1M tokens)$3.00 Input text (per 1M tokens)$0.75 Output audio (per 1M tokens)$12.00 Output text (per 1M tokens)$4.50

Usage

const response = await env.AI.run(
  'google/gemini-3.1-flash-tts',
  { text: 'Hello, welcome to Cloudflare AI Gateway!' },
)
console.log(response)

curl https://api.cloudflare.com/client/v4/accounts/$CLOUDFLARE_ACCOUNT_ID/ai/run \
  --header "Authorization: Bearer $CLOUDFLARE_API_TOKEN" \
  --header "Content-Type: application/json" \
  --data '{
  "model": "google/gemini-3.1-flash-tts",
  "input": {
    "text": "Hello, welcome to Cloudflare AI Gateway!"
  }
}'

{
  "audio": "data:audio/l16;base64,...",
  "gatewayMetadata": {
    "keySource": "Unified"
  }
}

Examples

Custom Voice — Generate speech with a specific voice

const response = await env.AI.run(
  'google/gemini-3.1-flash-tts',
  { text: 'The quick brown fox jumps over the lazy dog.', voice: 'Puck' },
)
console.log(response)

curl https://api.cloudflare.com/client/v4/accounts/$CLOUDFLARE_ACCOUNT_ID/ai/run \
  --header "Authorization: Bearer $CLOUDFLARE_API_TOKEN" \
  --header "Content-Type: application/json" \
  --data '{
  "model": "google/gemini-3.1-flash-tts",
  "input": {
    "text": "The quick brown fox jumps over the lazy dog.",
    "voice": "Puck"
  }
}'

{
  "audio": "data:audio/l16;base64,...",
  "gatewayMetadata": {
    "keySource": "Unified"
  }
}

Longer Text — Convert longer text to speech

const response = await env.AI.run(
  'google/gemini-3.1-flash-tts',
  {
    text: 'Artificial intelligence has transformed the way we interact with technology. From voice assistants to autonomous vehicles, AI is reshaping our daily lives and creating new possibilities for innovation.',
    voice: 'Charon',
  },
)
console.log(response)

curl https://api.cloudflare.com/client/v4/accounts/$CLOUDFLARE_ACCOUNT_ID/ai/run \
  --header "Authorization: Bearer $CLOUDFLARE_API_TOKEN" \
  --header "Content-Type: application/json" \
  --data '{
  "model": "google/gemini-3.1-flash-tts",
  "input": {
    "text": "Artificial intelligence has transformed the way we interact with technology. From voice assistants to autonomous vehicles, AI is reshaping our daily lives and creating new possibilities for innovation.",
    "voice": "Charon"
  }
}'

{
  "audio": "data:audio/l16;base64,...",
  "gatewayMetadata": {
    "keySource": "Unified"
  }
}

Narrative Voice — Generate speech with a narrative voice style

const response = await env.AI.run(
  'google/gemini-3.1-flash-tts',
  {
    text: 'Once upon a time, in a kingdom far away, there lived a brave knight who sought to protect the realm from all dangers.',
    voice: 'Kore',
  },
)
console.log(response)

curl https://api.cloudflare.com/client/v4/accounts/$CLOUDFLARE_ACCOUNT_ID/ai/run \
  --header "Authorization: Bearer $CLOUDFLARE_API_TOKEN" \
  --header "Content-Type: application/json" \
  --data '{
  "model": "google/gemini-3.1-flash-tts",
  "input": {
    "text": "Once upon a time, in a kingdom far away, there lived a brave knight who sought to protect the realm from all dangers.",
    "voice": "Kore"
  }
}'

{
  "audio": "data:audio/l16;base64,...",
  "gatewayMetadata": {
    "keySource": "Unified"
  }
}

Parameters

text

stringrequiredmaxLength: 10000The text to convert to speech. Maximum 10,000 characters.

voice

stringenum: Zephyr, Puck, Charon, Kore, Fenrir, Leda, Orus, Aoede, Callirrhoe, Autonoe, Enceladus, Iapetus, Umbriel, Algieba, Despina, Erinome, Algenib, Rasalgethi, Laomedeia, Achernar, Alnilam, Schedar, Gacrux, Pulcherrima, Achird, Zubenelgenubi, Vindemiatrix, Sadachbia, Sadaltager, SulafatThe voice to use for speech synthesis

temperature

numberminimum: 0maximum: 2Controls randomness in generation (0-2)

topP

numberminimum: 0maximum: 1Nucleus sampling threshold (0-1). Tokens with cumulative probability up to topP are considered

topK

integerexclusiveMinimum: 0maximum: 9007199254740991Only sample from the top K tokens. Smaller K = more focused, larger K = more diverse

maxOutputTokens

integerexclusiveMinimum: 0maximum: 9007199254740991Maximum number of tokens to generate

▶stopSequences[]

arraySequences where the model will stop generating further tokens

audio

stringBase64-encoded audio data (WAV format)

API Schemas (Raw)

Input

Output