01 · Studio

Generate speech that
sounds human.

Type a script, pick from 281 real Inworld voices, and steer tone with inline emotion tags or natural-language direction. ⌘⏎ to generate.

01Script

0 / 2000

Emotion tags click to insert before the phrase it should affect

Non-verbal sounds dropped exactly where they occur

Pauses insert a silence at the cursor (SSML break)

Global direction natural-language tone, best with tts-2 (optional)

02Voice

of

Model

Delivery tts-2

Speed 1.0×

Variation 1.0

Ready to generate

0:00 0:00

02 · Voices

281 real voices,
browse and filter.

Every voice below comes straight from the Inworld catalog pick one here and it becomes the active voice in the Studio.

Clone your own voice

5–15s sample · wav/mp3 · ≤4MB

Voice name

Language

Audio sample — record or upload

0s

Transcription (optional, improves quality)

All voices

of

03 · Costs

What this
actually costs.

Real published rates, and a running total for what you've generated this session nothing here is estimated after the fact.

This session · generations

0

This session · characters billed

0

This session · est. cost

$0.00

This session · enhance calls

0

Inworld TTS per character

Realtime TTS-2

$5 / 1M chars

Newest, most expressive, natural-language steering.

TTS-1.5 Max default

$10 / 1M chars

Best quality-to-price; what this app uses by default.

TTS-1.5 Mini

$15→7 / 1M chars

Cheapest/fastest at volume; tiered pricing.

Quick calculator

~1,000 characters ≈ ~165 words ≈ ~1 minute of speech. Rates as of Aug 2026 see COSTING.md for the full breakdown and OpenAI enhance pricing.

Generate speech thatsounds human.