Last updated 07-27-2026
Category:
Reviews:
Join thousands of AI enthusiasts in the World of AI!
Inworld
Inworld builds realtime voice infrastructure for developers who need speech-to-speech agents, not just a TTS API bolted onto a chatbot. The platform spans text-to-speech, speech-to-text, LLM routing, and a unified Realtime API that chains those layers in one WebSocket session. Products ship through REST and WebSocket endpoints at api.inworld.ai, with a browser portal for keys, billing, and playgrounds.
Where ElevenLabs and Cartesia mostly sell standalone TTS, Inworld bets on the full voice stack. Its STT layer extracts five paralinguistic signals per audio chunk (emotion, age, accent, pitch, vocal style) and feeds them into the LLM and TTS downstream. The Realtime Router routes across 220+ models at provider cost with zero markup, which is unusual compared to gateways that charge a 5% fee on credits.
Game studios, customer support teams, and indie developers use Inworld for voice agents, interactive NPCs, language tutoring apps, and AI companions. The On-Demand tier includes up to 70 minutes of TTS free, which lowers the barrier for prototyping before committing to a $25 per month Creator plan.
Enterprise teams get SOC 2 Type II, HIPAA, and GDPR compliance options, with Zero Data Retention available as an add-on on Growth plans. That matters for healthcare intake and regulated industries where voice data cannot linger on vendor servers.
Realtime TTS-2 delivers sub-200ms first-chunk latency with inline steering tags like [Speak softly] and non-verbals like [sigh]
Instant voice cloning from 5 to 15 seconds of audio, with cross-lingual delivery across 100+ languages on TTS-2
Realtime Router provides access to 220+ LLM models with zero markup on provider rates
Realtime STT streams transcription down to $0.10 per hour with five voice profile signals per audio chunk
Realtime API runs STT, LLM, and TTS in one WebSocket with sub-second speech-to-speech latency
On-Demand tier includes up to 70 minutes of TTS and 100 custom voices at no cost
Full voice stack (TTS, STT, routing, speech-to-speech) in one platform instead of wiring separate vendors
LLM Router passes through 220+ models at provider rates with zero markup
On-Demand tier includes up to 70 minutes of TTS free for prototyping
Production tiers scale quickly, with the Developer plan at $300 per month before usage overages
Professional voice cloning requires contacting sales or upgrading to Developer or Growth
Built for developers via API and SDKs, with no no-code builder for non-technical users
Is Inworld free to use?
Yes. Inworld's On-Demand plan starts free and includes up to 70 minutes of TTS, 100 custom voices, voice cloning, Realtime API access, and 220+ LLM models via Router. Paid plans begin at $25 per month on the Creator tier.
What does Inworld's Realtime API do?
Inworld's Realtime API runs speech-to-speech in one WebSocket session, chaining STT voice profiling, LLM reasoning, and Realtime TTS-2 output. It supports hundreds of LLM models, semantic turn detection, tool calling, and sub-second end-to-end latency.
How much does Inworld TTS cost?
Inworld Realtime TTS-2 starts at $25 per million characters on the On-Demand plan. Rates fall to $15 per million at the $300 per month Developer tier and $12.50 at the $1,500 per month Growth tier. Enterprise pricing can reach $10 per million characters.
Which LLM models does Inworld Router support?
Inworld Router provides access to 220+ models from providers including OpenAI, Anthropic, Google, xAI, Groq, and Mistral. Routed third-party models carry zero markup, and the Router supports CEL-based conditional routing and A/B testing.
Does Inworld support voice cloning?
Yes. Inworld offers instant voice cloning from 5 to 15 seconds of audio on all plans. Professional voice cloning, which uses 30+ minutes of clean audio for higher fidelity, is available as an add-on on Developer plans and included on Growth.
What compliance certifications does Inworld have?
Inworld is SOC 2 Type II certified and offers HIPAA and GDPR compliance. Zero Data Retention is available as an add-on on Growth plans, and Enterprise customers can access EU and India data residency options.