Inworld Realtime TTS 2.0
Inworld Realtime TTS 2.0 (inworld-tts-2) is Inworld’s most expressive text-to-speech model. It builds on Realtime TTS 1.5 with expressive speech and real-time latency, and adds natural-language steering and better support for languages other than English.
What’s new in 2.0
- Natural-language steering. Direct any voice with bracketed natural-language cues like
[say excitedly],[whisper in a hushed style], or[say sadly with deliberate pauses in a low voice]. There’s no preset list. Write your directions in plain language. - Better multilingual speech. More natural, native speech across 15 production languages, plus experimental support for 90+ additional languages.
- Better alphanumerics. It reads phone numbers, SKUs, and codes more accurately.
- Quality. It builds on Realtime TTS 1.5 Max, which ranked first on Artificial Analysis, with a lower word error rate, tighter alignment, and more natural speech.
Steering with natural language
Wrap your direction in square brackets and place it before the text it applies to. Think of it as giving direction to a voice actor.
Free-form direction
[speak as if barely holding back rage, forcing every word through gritted teeth] I have told you. Repeatedly. And you STILL didn't listen.
[overwhelmed with excitement and barely able to contain yourself] We just hit a million users. I still can't believe it!
[slow and hushed with every word weighted by grief] I got the call this morning. He's gone.
Single-property direction
| Quality | Examples |
|---|---|
| Articulation | [say with force], [articulate clearly], [say with deliberate pauses] |
| Intonation | [say with a falling pitch], [say with a rising pitch] |
| Volume | [very loud], [very quiet] |
| Pitch | [say in a low tone], [say in a high pitch] |
| Range | [say playfully], [say with no pitch variation] |
| Speed | [very fast], [very slow] |
| Vocal style | [whisper in a hushed style], [sing joyfully], [give a nasal quality] |
The more you describe how you want the voice to perform, the better the output. A bare tag like [sad] gives the model one dimension to work with. A fuller instruction like [say sadly with deliberate pauses in a low voice and hushed style] combines mood, rhythm, pitch, and mode, and gives a more convincing performance.
Inline non-verbals
Insert human sounds at any point in the text:
[laugh] [breathe] [clear throat] [sigh] [cough] [yawn]
I told him what happened, and he just [laugh] couldn't believe it!
Emphasis
Capitalize letters within your input text to draw attention to specific words or syllables.
I told you NOT to open that door.
Are you seriously asking if I want pizza? AbsoLUTEly I do.
Languages
Production languages (best quality): English, Chinese, Japanese, Korean, Russian, Italian, Spanish, Portuguese, French, German, Polish, Dutch, Hindi, Hebrew, Arabic.
Experimental support for 90+ additional languages including Vietnamese, Thai, Turkish, Indonesian, Swedish, and many more. See the language support docs for the full list.
Set language to auto (default) to let the model detect the language, or pick one of the production codes for best results.
Voices
| Voice | Description |
|---|---|
Ashley |
A warm, natural female voice |
Dennis |
Middle-aged man with a smooth, calm and friendly voice |
Alex |
Energetic and expressive mid-range male voice, with a mildly nasal quality |
Darlene |
Soothing, comforting Southern female voice, ideal for bedtime stories and narrations |
Set voice_id to one of these preset voices. Custom voices you clone in your own Inworld account are scoped to that account’s API key and aren’t usable here.
Best practices
- Match the instruction to the text. Mismatches like
[say sadly]on celebratory text degrade output quality. - Avoid conflicting instructions. Opposing directions in the same tag, such as
[whisper in a hushed style]with[very loud], give unpredictable results. - Place steering tags before the text they apply to. Non-verbal tags like
[laugh]are the exception. Put them inline where the sound should happen.
Choosing between Inworld TTS models
- Realtime TTS 2.0 (inworld/realtime-tts-2). Most expressive, with natural-language steering and stronger multilingual support. Best when you need fine-grained control over delivery.
- Realtime TTS 1.5 Max (inworld/realtime-tts-1.5-max). Ranked first on Artificial Analysis, with latency under 200 ms. Best for raw 1.5-era quality.
- Realtime TTS 1.5 Mini (inworld/realtime-tts-1.5-mini). The fastest, at about 120 ms, and the lowest cost. Best for high-volume, latency-sensitive applications.