--- title: Inworld TTS 2 description: Inworld's most powerful and expressive text-to-speech model. Builds on TTS 1.5 with rich expressive speech, real-time latency, natural language steering (e.g. [whisper], [say excitedly]), and stronger multilingual support across 15 production languages plus 90+ experimental languages. image: https://developers.cloudflare.com/dev-products-preview.png --- > Documentation Index > Fetch the complete documentation index at: https://developers.cloudflare.com/ai/llms.txt > Use this file to discover all available pages before exploring further. [Skip to content](#%5Ftop) ![Inworld logo](https://developers.cloudflare.com/_astro/inworld.BDwMAXI2.svg) # Inworld TTS 2 Text-to-Speech • Inworld • Proxied `inworld/tts-2` Inworld's most powerful and expressive text-to-speech model. Builds on TTS 1.5 with rich expressive speech, real-time latency, natural language steering (e.g. \[whisper\], \[say excitedly\]), and stronger multilingual support across 15 production languages plus 90+ experimental languages. | Model Info | | | ----------------- | --------------------------------------------------------------------------------------------------------------- | | Terms and License | [link ↗](https://inworld.ai/terms) | | More information | [link ↗](https://docs.inworld.ai/tts/realtime-tts-2-preview) | | Pricing | [View pricing in the Cloudflare dashboard ↗](https://dash.cloudflare.com/?to=/:account/ai/models/inworld/tts-2) | ## Usage TypeScript ``` const response = await env.AI.run( 'inworld/tts-2', { text: 'Hello! Welcome to Cloudflare AI Gateway. Let me show you what we can do.', voice_id: 'Dennis', output_format: 'mp3', temperature: 1, timestamp_type: 'none', }, { gateway: { id: 'default' }, } ) console.log(response) ``` * [ Output ](#tab-panel-302) * [ Raw response ](#tab-panel-303) ``` { "state": "Completed", "result": { "audio": "https://pub-04a6d208d361438ea01b797e6973bd19.r2.dev/catalog/inworld__tts-2/simple-speech.mp3" }, "gatewayMetadata": { "keySource": "Unified" } } ``` ## Examples **Natural Language Steering** — Direct the voice with bracketed natural-language cues for emotion, pace, and style. TypeScript ``` const response = await env.AI.run( 'inworld/tts-2', { text: "[speak with excitement] I'm really excited about Inworld's new model. Have you tried out the steering capabilities? It's pretty cool!", voice_id: 'Dennis', output_format: 'mp3', temperature: 1, timestamp_type: 'none', }, { gateway: { id: 'default' }, } ) console.log(response) ``` * [ Output ](#tab-panel-304) * [ Raw response ](#tab-panel-305) ``` { "state": "Completed", "result": { "audio": "https://pub-04a6d208d361438ea01b797e6973bd19.r2.dev/catalog/inworld__tts-2/natural-language-steering.mp3" }, "gatewayMetadata": { "keySource": "Unified" } } ``` **Whisper** — Use steering tags to whisper TypeScript ``` const response = await env.AI.run( 'inworld/tts-2', { text: '[whisper] This is a secret just between us.', voice_id: 'Dennis', output_format: 'mp3', temperature: 1, timestamp_type: 'none', }, { gateway: { id: 'default' }, } ) console.log(response) ``` * [ Output ](#tab-panel-306) * [ Raw response ](#tab-panel-307) ``` { "state": "Completed", "result": { "audio": "https://pub-04a6d208d361438ea01b797e6973bd19.r2.dev/catalog/inworld__tts-2/whisper.mp3" }, "gatewayMetadata": { "keySource": "Unified" } } ``` **High Quality Audio** — Higher sample rate for studio quality TypeScript ``` const response = await env.AI.run( 'inworld/tts-2', { text: 'This recording is generated at studio quality for the best possible listening experience.', voice_id: 'Dennis', output_format: 'mp3', sample_rate: 48000, temperature: 1, timestamp_type: 'none', }, { gateway: { id: 'default' }, } ) console.log(response) ``` * [ Output ](#tab-panel-308) * [ Raw response ](#tab-panel-309) ``` { "state": "Completed", "result": { "audio": "https://pub-04a6d208d361438ea01b797e6973bd19.r2.dev/catalog/inworld__tts-2/high-quality-audio.mp3" }, "gatewayMetadata": { "keySource": "Unified" } } ``` **With Text Normalization** — Expand numbers and abbreviations before synthesis TypeScript ``` const response = await env.AI.run( 'inworld/tts-2', { text: 'The meeting is at 3:30 PM on Jan 15th, 2026. Please confirm by calling 555-0123.', voice_id: 'Dennis', output_format: 'mp3', temperature: 1, timestamp_type: 'none', apply_text_normalization: true, }, { gateway: { id: 'default' }, } ) console.log(response) ``` * [ Output ](#tab-panel-310) * [ Raw response ](#tab-panel-311) ``` { "state": "Completed", "result": { "audio": "https://pub-04a6d208d361438ea01b797e6973bd19.r2.dev/catalog/inworld__tts-2/with-text-normalization.mp3" }, "gatewayMetadata": { "keySource": "Unified" } } ``` ## Parameters * [ Input ](#tab-panel-312) * [ Output ](#tab-panel-313) text `string`requiredmaxLength: 2000The text to be synthesized into speech. Maximum input of 2,000 characters. voice\_id `string`requireddefault: Dennisenum: Loretta, Darlene, Marlene, Hank, Evelyn, Celeste, Pippa, Tessa, Liam, Callum, Hamish, Abby, Graham, Rupert, Mortimer, Snik, Anjali, Saanvi, Arjun, Claire, Oliver, Simon, Elliot, James, Serena, Gareth, Vinny, Lauren, Jessica, Ethan, Tyler, Jason, Chloe, Veronica, Victoria, Miranda, Sebastian, Victor, Malcolm, Nate, Brian, Amina, Kelsey, Derek, Evan, Kayla, Jake, Grant, Tristan, Nadia, Selene, Marcus, Riley, Damon, Cedric, Mia, Naomi, Jonah, Levi, Avery, Brandon, Conrad, Bianca, Lucian, Trevor, Alex, Ashley, Craig, Deborah, Dennis, Edward, Elizabeth, Hades, Julia, Pixie, Mark, Olivia, Priya, Ronald, Sarah, Shaun, Theodore, Timothy, Wendy, Dominus, Hana, Clive, Carter, Blake, Luna, Reed, Duncan, Felix, Eleanor, SophieThe ID of the voice to use for synthesizing speech. Defaults to Dennis. output\_format `string`requireddefault: mp3enum: mp3, opus, wav, flacThe output format for the audio. Supported formats are mp3, opus, wav, and flac. Defaults to mp3. bit\_rate `integer`minimum: \-9007199254740991maximum: 9007199254740991Bits per second of the audio. Only for compressed audio formats (mp3, opus). The default is 128,000. sample\_rate `integer`minimum: \-9007199254740991maximum: 9007199254740991The synthesis sample rate in hertz. Accepts: 8000, 16000, 22050, 24000, 32000, 44100, 48000\. The default is 48,000. speaking\_rate `number`minimum: 0.5maximum: 1.5Speaking rate/speed, in the range \[0.5, 1.5\]. The default is 1.0\. We recommend using values above 0.8 to ensure high quality. temperature `number`requireddefault: 1minimum: 0.01maximum: 2Determines the degree of randomness when sampling audio tokens. Defaults to 1.0\. Accepts values between 0 (exclusive) and 2 (inclusive). Higher values = more expressive, lower values = more deterministic. timestamp\_type `string`requireddefault: noneenum: none, word, characterControls timestamp metadata returned with the audio. "word" returns word-level timing, "character" returns character-level timing. Note: adds latency. Defaults to none. apply\_text\_normalization `boolean`When enabled, text normalization expands numbers, dates, times, and abbreviations before converting to speech. Turning this off may reduce latency. audio `string`URL to the generated audio file ## API Schemas (Raw) Input Output ```json {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"item":{"@id":"/directory/","name":"Directory"}},{"@type":"ListItem","position":2,"item":{"@id":"/ai/","name":"AI"}},{"@type":"ListItem","position":3,"item":{"@id":"/ai/models/","name":"Models"}}]} ```