Last updated 08-14-2026
Category:
Reviews:
Join thousands of AI enthusiasts in the World of AI!
Inworld AI
Inworld AI sells realtime voice APIs for text-to-speech, speech-to-text, speech-to-speech agents, and LLM routing, including avatar and NPC voice generation for games. Its Realtime TTS-2 models advertise 100 ms time-to-first-byte, voice cloning from 15 seconds of audio, bracketed voice steering tags, and support for 200+ languages on a single voice.
Compared with stitching together separate TTS, STT, and gateway vendors, Inworld bundles those layers into Realtime API sessions and a zero-markup Router across 220+ LLM models. The homepage publishes side-by-side rate cuts, such as $12.50 per 1M characters for Realtime TTS versus $100 at ElevenLabs on comparable growth plans, which targets teams shipping consumer voice apps at scale.
Game studios still use Inworld for NPC and companion voices, but the current product also targets education, health, social apps, and agentic workforce tools. Developers sign up on platform.inworld.ai, pick an On-Demand free tier or a paid monthly plan with credits, and integrate through REST or WebSocket APIs documented at docs.inworld.ai.
Realtime TTS-2 advertises 100 ms TTFB; TTS-2 Flash advertises 20 ms
Voice cloning from 15 seconds of audio across 15 supported languages
Realtime API runs STT, LLM steering, and TTS in one WebSocket session
Router provides zero-markup access to 220+ LLM models from major providers
Realtime STT streams with built-in emotion, age, accent, pitch, and style signals
Bracketed inline tags steer tone, speed, volume, and pauses in generated speech
SOC 2 Type II, HIPAA, and GDPR compliance options on higher tiers
Bundles TTS, STT, speech-to-speech, and LLM routing instead of wiring multiple vendors.
Published rate comparisons show steep discounts versus standalone TTS and gateway providers.
Voice steering tags and profiling signals give fine control over delivery and user context.
Pricing tiers and credit math take time to model for high-volume production apps.
Advanced compliance features such as HIPAA and zero data retention require Growth add-ons or Enterprise.
The platform is API-first; non-developers need engineering help to ship a finished voice product.
What does Inworld AI offer?
Inworld AI sells realtime voice APIs for text-to-speech, speech-to-text, speech-to-speech agents, and LLM routing. Developers integrate through platform.inworld.ai with docs at docs.inworld.ai.
Is there a free Inworld AI plan?
Yes. Inworld AI's On-Demand plan starts free and includes up to 70 minutes of TTS usage, 100 custom voices, voice cloning, Realtime API access, and community support for evaluation and prototyping.
How much does Inworld AI cost?
Inworld AI paid plans start at $25 per month for Creator, $100 for Builder, $300 for Developer, and $1,500 for Growth, with monthly credits included. On-Demand pay-as-you-go rates include $25 per 1M characters for Realtime TTS-2.
Which languages does Inworld TTS support?
Inworld advertises 200+ languages for Realtime TTS, including English, Spanish, French, Korean, Chinese, Hindi, Japanese, and German. Cross-lingual voice cloning lets one cloned voice speak multiple languages.
Does Inworld AI create 3D avatars?
No. Inworld AI powers character brains and voice output; visuals are up to you. Its site notes that demo avatars use Ready Player Me, but the platform is model-agnostic for appearance.
What is the Inworld Realtime API?
The Inworld Realtime API is an end-to-end speech-to-speech stack with custom voices, tool calling, turn detection, and provider-agnostic LLM routing over one WebSocket connection.