Inworld AI provides realtime voice models (text to speech, speech to text, speech to speech) and hosted open LLM inference for developers building consumer apps.
About
Inworld AI is a platform of realtime voice models and hosted open large language models aimed at consumer-facing applications. It addresses the problem of adding natural, low-latency speech and language capabilities to apps such as AI companions, tutors, wellness coaches, voice agents for support and sales, and interactive games and media, without a team having to run its own model infrastructure.
The company offers speech models (Realtime TTS-2 and TTS-2 Flash for text to speech, Realtime STT-1 for speech to text, and a speech-to-speech product for full voice conversations) plus Inference, an OpenAI-compatible API serving open LLMs on a fleet Inworld operates. Each model has its own API and they can be combined; there is also a Compute option for dedicated deployments. Inworld states its first audio can arrive in as little as 50 ms and that it serves tens of trillions of tokens a month.
Key Features
Pricing & Fees
- Pricing: Plans are credit-based and start with a free On-Demand tier (TTS-2 at $25/1M chars, TTS-2 Flash at $15/1M chars, STT-1 at $0.15/hr, LLM inference from $0.07/1M tokens). Paid tiers are Creator at $25/mo, Builder at $100/mo, Developer at $300/mo and Growth at $1,500/mo, each adding monthly credits and discounts of up to roughly 47% on TTS and STT rates. Enterprise-scale TTS-2 Flash drops to $5/1M chars.
- Free tier: The On-Demand plan costs $0 with pay-as-you-go rates, up to 70 minutes of TTS included, 100 custom voices and 2 professional voice clones.
- Integrations: Inworld voices work inside LiveKit Agents without a second API key, and the company publishes a "skill" that can be pasted into coding agents such as Claude Code or Cursor.
- Reliability and scale: Inworld reports 99.999% of requests served for its Priority LLM and voice models, with fewer than 1 in 100,000 requests failing over a 30-day measurement window.
Who It's For
Inworld AI is built for engineering teams creating consumer applications that people talk to every day. That includes developers of AI companion and social apps, language tutors and education products, health and fitness coaches, voice agents for customer support and sales operations, and studios making games and narrated media that need live dialogue and expressive character voices. The OpenAI-compatible inference API also suits teams that want open LLMs served by a third party rather than managing their own GPU fleet.
FAQ
What is Inworld AI?Inworld AI is a platform offering realtime speech models (text to speech, speech to text and speech to speech) and an OpenAI-compatible inference service for open LLMs, built for consumer-facing applications.
Does Inworld AI have a free plan?Yes, the On-Demand plan costs $0 on a pay-as-you-go basis and includes up to 70 minutes of TTS, 100 custom voices and 2 professional voice clones.
Which platforms and integrations does Inworld AI support?It offers REST, WebSocket and WebRTC APIs, works with the OpenAI SDK by changing the base URL, integrates with LiveKit Agents without a second API key, and provides a skill for coding agents like Claude Code and Cursor.
How many languages does Inworld AI's text to speech support?Realtime TTS-2 and TTS-2 Flash speak about 500 languages and dialects from a single voice model, while Realtime STT-1 transcribes 30 languages.
What about privacy and compliance?Inworld lists SOC 2 Type II, HIPAA and GDPR compliance, with voice infrastructure in four regions and LLM serving in the US; SOC 2 Type II and GDPR coverage applies to every pricing plan.
Related Tutorials
View all →
Related Tools
View all →
FunASR - End-to-End Speech Recognition Toolkit Industrial-grade speech recognition toolkit. 170x realtime on GPU, 50+ languages, speaker diarization, emotion detection, OpenAI-compatible API, vLLM acceleration, and agent integration. 
AI Translator - Accurate Translation AI Translator provides high-accuracy translations with 98% precision, supporting 1200+ AI models and 100+ languages. 
IBM Watson Text-to-Speech Convert text to natural speech using IBM Watson's API and SaaS. 
NaturalReader NaturalReader is an AI text-to-speech platform that converts documents, webpages and scanned text into natural-sounding audio for personal, commercial and educational use. 
VoxCPM - Voice Cloning & Speech Synthesis VoxCPM generates natural, expressive speech from text. It leverages advanced LLMs to create realistic voiceovers for various applications.