- SI Score
- 80/100
- Pricing
- Not confirmed
- Free plan
- Not confirmed
- Sign-up required
- Not confirmed
Ask SI about Inworld AI
Answers use saved listing facts and recorded sources.Homepage screenshot
What it does
Inworld AI provides speech and language models for consumer-facing apps. The site lists realtime text to speech, speech to text and speech to speech, plus hosted open models for text, image, video and decision tasks through one API key. Realtime TTS-2, Realtime TTS-2 Flash and Realtime STT-1 are described with latency figures, language coverage and pricing. Developers can get an API key, open a playground, read documentation and quickstarts, install a CLI, and reach sales. Voice and Inference are the two product areas covered in the docs.
The home page states first audio as low as 50 ms for Realtime TTS-2 Flash and under 70 ms to a final transcript for Realtime STT-1. Realtime TTS-2 is described as supporting about 500 languages and dialects. Speech to speech is described as listening, reasoning and talking back on one WebSocket or WebRTC connection, with a stated 204 ms from turn closed to first audio. Hosted models listed include GLM 5.3 Flash, Gemma 4 31B, Gemma 4 26B, MiMo 2.6 Pro, DeepSeek V4.1 Flash and Clef, with per-million-token input and output prices. Serving regions named are US Central Iowa, US East Virginia, Europe the Netherlands, India Mumbai and Asia Pacific Hong Kong. The docs describe Voice and Inference products, quickstarts, an API reference, SDKs and a CLI. Links go to pricing, contact sales, docs and platform signup.