Overview
Inworld AI provides a scalable infrastructure for building expressive, real-time characters and voice agents. The platform focuses on ultra-low latency TTS, instant voice cloning, multimodal runtime pipelines, and integrations (Portal/API/on-prem/on-device). It supports multilingual TTS models, safety & memory features, and a marketplace of hosted and on-prem model deployments. Inworld promotes open-source research and training code for its TTS models and highlights use cases across games, media, contact centers, and training simulations.
Key Features
Real-time TTS & low-latency streaming
Delivers sub-250 ms latency streaming speech optimized for conversational agents.
Instant voice cloning
Zero-shot cloning from 2–15 seconds of audio and professional cloning for high-fidelity voices.
Multimodal runtime pipelines
Runtime pipelines for characters that integrate voice, memory, knowledge, and behavior graphs.
Expressive controls (emotion, delivery, non-verbal sounds)
Voice tags and audio markups add emotion, delivery style, and non-verbal audio cues.
On-prem & enterprise deployments
Options for on-prem, hosted, or on-device deployments and contact-for-pricing enterprise models.
Safety, Memory & Knowledge integration
Built-in governance features (safety policies), memory, and knowledge modules included.
Who Can Use This Tool?
- Developers:Integrate real-time expressive voices and character runtimes into apps and games.
- Game Studios:Build scalable, interactive characters and multiplayer voice experiences for players.
- Contact Center Teams:Deploy voice agents and CX experiences with lower latency and improved CSAT.
- Training & Education:Create immersive simulations and role-play scenarios with expressive AI characters.
Pricing Plans
Inworld text-to-speech priced per 1M characters, usage-based.
- ✓Provider: Inworld
- ✓Cost: $5 / 1M characters
- ✓Approx: ~$0.005 / minute
- ✓On-prem available (contact for pricing)
Higher-quality Inworld TTS tier, usage-priced per characters.
- ✓Provider: Inworld
- ✓Cost: $10 / 1M characters
- ✓Approx: ~$0.01 / minute
- ✓On-prem available (contact for pricing)
Safety features included with Inworld models; no additional usage charge.
- ✓Provider: Inworld
- ✓Input cost: Included
- ✓Output cost: Included
- ✓Included with platform usage
Memory capabilities included with Inworld models at no charge.
- ✓Provider: Inworld
- ✓Input cost: Included
- ✓Output cost: Included
- ✓Included with platform usage
Knowledge features included with Inworld models at no extra cost.
- ✓Provider: Inworld
- ✓Input cost: Included
- ✓Output cost: Included
- ✓Included with platform usage
On-prem deployment of gpt-oss 20B — contact Inworld for pricing.
- ✓Provider: Inworld (On-prem)
- ✓Input cost: -
- ✓Output cost: Contact for pricing
- ✓On-prem deployment
On-prem gemma3 12B model — contact for pricing and deployment details.
- ✓Provider: Inworld (On-prem)
- ✓Input cost: -
- ✓Output cost: Contact for pricing
- ✓On-prem deployment
On-prem gemma3 27B — contact Inworld for pricing and availability.
- ✓Provider: Inworld (On-prem)
- ✓Input cost: -
- ✓Output cost: Contact for pricing
- ✓On-prem deployment
On-prem Llama 3.1 8B deployment with contact-for-pricing.
- ✓Provider: Inworld (On-prem)
- ✓Input cost: -
- ✓Output cost: Contact for pricing
- ✓On-prem deployment
On-prem voice activity detection included with deployment, no usage charge listed.
- ✓Provider: Inworld (On-prem)
- ✓Input cost: Included
- ✓Output cost: Included
- ✓On-prem feature
Anthropic model charged per 1M tokens for input and output.
- ✓Provider: Anthropic
- ✓Input cost: $15 / 1M tokens
- ✓Output cost: $75 / 1M tokens
Anthropic Opus 4.1 model charged per 1M tokens for input and output.
- ✓Provider: Anthropic
- ✓Input cost: $15 / 1M tokens
- ✓Output cost: $75 / 1M tokens
Anthropic Sonnet model, low-cost option billed per 1M tokens.
- ✓Provider: Anthropic
- ✓Input cost: $3 / 1M tokens
- ✓Output cost: $15 / 1M tokens
Anthropic Sonnet 3.7 model billed per 1M tokens.
- ✓Provider: Anthropic
- ✓Input cost: $3 / 1M tokens
- ✓Output cost: $15 / 1M tokens
Anthropic Haiku model charged per 1M tokens.
- ✓Provider: Anthropic
- ✓Input cost: $0.25 / 1M tokens
- ✓Output cost: $1.25 / 1M tokens
Anthropic 3.5 Haiku model charged per 1M tokens.
- ✓Provider: Anthropic
- ✓Input cost: $0.8 / 1M tokens
- ✓Output cost: $4 / 1M tokens
Google Gemini 2.5 Flash accessed through Vertex AI; billed per 1M tokens.
- ✓Provider: Google
- ✓Input cost: $0.3 / 1M tokens
- ✓Output cost: $2.5 / 1M tokens
- ✓Access through Google / Vertex AI
Gemini 2.5 Pro pricing for <=200K input tokens via Vertex AI.
- ✓Provider: Google (Through Vertex AI)
- ✓Input cost: $1.25 / 1M tokens (<=200K input tokens)
- ✓Output cost: $10 / 1M tokens
Gemini 2.5 Pro pricing for >200K input tokens via Vertex AI.
- ✓Provider: Google (Through Vertex AI)
- ✓Input cost: $2.5 / 1M tokens (>200K input tokens)
- ✓Output cost: $15 / 1M tokens
Gemini 2.5 Flash Lite via Vertex AI billed per 1M tokens.
- ✓Provider: Google (Through Vertex AI)
- ✓Input cost: $0.1 / 1M tokens
- ✓Output cost: $0.4 / 1M tokens
Gemini 2.0 Flash via Vertex AI billed per 1M tokens.
- ✓Provider: Google (Through Vertex AI)
- ✓Input cost: $0.1 / 1M tokens
- ✓Output cost: $0.4 / 1M tokens
Gemini 2.0 Flash-Lite priced per 1M tokens.
- ✓Input cost: $0.075 / 1M tokens
- ✓Output cost: $0.3 / 1M tokens
OpenAI realtime model priced per 1M tokens.
- ✓Provider: OpenAI
- ✓Input cost: $4 / 1M tokens
- ✓Output cost: $16 / 1M tokens
gpt-5 model billed per 1M tokens.
- ✓Provider: OpenAI
- ✓Input cost: $1.25 / 1M tokens
- ✓Output cost: $10 / 1M tokens
gpt-5-mini billed per 1M tokens.
- ✓Provider: OpenAI
- ✓Input cost: $0.25 / 1M tokens
- ✓Output cost: $2 / 1M tokens
gpt-5-nano billed per 1M tokens.
- ✓Provider: OpenAI
- ✓Input cost: $0.05 / 1M tokens
- ✓Output cost: $0.4 / 1M tokens
Latest gpt-5 chat model priced per 1M tokens.
- ✓Input cost: $1.25 / 1M tokens
- ✓Output cost: $10 / 1M tokens
gpt-4.1 model priced per 1M tokens.
- ✓Input cost: $2 / 1M tokens
- ✓Output cost: $8 / 1M tokens
GPT-4.1 mini priced per 1M tokens.
- ✓Input cost: $0.4 / 1M tokens
- ✓Output cost: $1.6 / 1M tokens
GPT-4.1 nano priced per 1M tokens.
- ✓Input cost: $0.1 / 1M tokens
- ✓Output cost: $0.4 / 1M tokens
GPT-4o model priced per 1M tokens.
- ✓Input cost: $2.5 / 1M tokens
- ✓Output cost: $10 / 1M tokens
gpt-4o (2024-05-13) priced per 1M tokens.
- ✓Input cost: $5 / 1M tokens
- ✓Output cost: $15 / 1M tokens
GPT-4o-mini priced per 1M tokens.
- ✓Input cost: $0.15 / 1M tokens
- ✓Output cost: $0.6 / 1M tokens
o1 model priced per 1M tokens.
- ✓Input cost: $15 / 1M tokens
- ✓Output cost: $60 / 1M tokens
o1-pro model priced per 1M tokens.
- ✓Input cost: $150 / 1M tokens
- ✓Output cost: $600 / 1M tokens
o3-pro model priced per 1M tokens.
- ✓Input cost: $20 / 1M tokens
- ✓Output cost: $80 / 1M tokens
o3 model priced per 1M tokens.
- ✓Input cost: $2 / 1M tokens
- ✓Output cost: $8 / 1M tokens
o4-mini priced per 1M tokens.
- ✓Input cost: $1.1 / 1M tokens
- ✓Output cost: $4.4 / 1M tokens
o3-mini priced per 1M tokens.
- ✓Input cost: $1.1 / 1M tokens
- ✓Output cost: $4.4 / 1M tokens
o1-mini priced per 1M tokens.
- ✓Input cost: $1.1 / 1M tokens
- ✓Output cost: $4.4 / 1M tokens
Mistral Small 3.2 model priced per 1M tokens.
- ✓Provider: Mistral
- ✓Input cost: $0.1 / 1M tokens
- ✓Output cost: $0.3 / 1M tokens
Ministral 8B 24.10 model priced per 1M tokens.
- ✓Provider: Mistral
- ✓Input cost: $0.1 / 1M tokens
- ✓Output cost: $1 / 1M tokens
DeepSeek V3.1 model priced per 1M tokens.
- ✓Provider: Fireworks
- ✓Input cost: $0.56 / 1M tokens
- ✓Output cost: $1.68 / 1M tokens
Meta Llama 3.1 405B priced per 1M tokens.
- ✓Provider: Fireworks
- ✓Input cost: $3 / 1M tokens
- ✓Output cost: $3 / 1M tokens
Meta Llama 4 Maverick (Basic) priced per 1M tokens.
- ✓Input cost: $0.22 / 1M tokens
- ✓Output cost: $0.88 / 1M tokens
Meta Llama 3.2 3B Instruct priced per 1M tokens.
- ✓Input cost: $0.1 / 1M tokens
- ✓Output cost: $0.1 / 1M tokens
Meta Llama 3.1 8B Instruct priced per 1M tokens.
- ✓Input cost: $0.2 / 1M tokens
- ✓Output cost: $0.2 / 1M tokens
Meta Llama 3.3 70B Instruct priced per 1M tokens.
- ✓Input cost: $0.9 / 1M tokens
- ✓Output cost: $0.9 / 1M tokens
Qwen3 235B Family and GLM-4.5 Air priced per 1M tokens.
- ✓Input cost: $0.22 / 1M tokens
- ✓Output cost: $0.88 / 1M tokens
Kimi K2 Instruct priced per 1M tokens.
- ✓Input cost: $0.6 / 1M tokens
- ✓Output cost: $2.5 / 1M tokens
Qwen3 Coder 480B priced per 1M tokens.
- ✓Input cost: $0.45 / 1M tokens
- ✓Output cost: $1.8 / 1M tokens
OpenAI gpt OSS 120b (Fireworks) priced per 1M tokens.
- ✓Input cost: $0.15 / 1M tokens
- ✓Output cost: $0.6 / 1M tokens
OpenAI gpt OSS 20b (Fireworks) priced per 1M tokens.
- ✓Input cost: $0.07 / 1M tokens
- ✓Output cost: $0.3 / 1M tokens
GPT OSS 20B 128k (Groq) with 128k context billed per 1M tokens.
- ✓Input cost: $0.1 / 1M tokens
- ✓Output cost: $0.5 / 1M tokens
- ✓128k context
GPT OSS 120B 128k (Groq) with 128k context billed per 1M tokens.
- ✓Input cost: $0.15 / 1M tokens
- ✓Output cost: $0.75 / 1M tokens
- ✓128k context
Kimi K2 1T 256k (Groq) with 256k context billed per 1M tokens.
- ✓Input cost: $1 / 1M tokens
- ✓Output cost: $3 / 1M tokens
- ✓256k context
Llama 4 Scout (17Bx16E) 128k (Groq) priced per 1M tokens.
- ✓Input cost: $0.11 / 1M tokens
- ✓Output cost: $0.34 / 1M tokens
- ✓128k context
Llama 4 Maverick (17Bx128E) 128k (Groq) priced per 1M tokens.
- ✓Input cost: $0.2 / 1M tokens
- ✓Output cost: $0.6 / 1M tokens
- ✓128k context
Llama Guard 4 12B 128k (Groq) priced per 1M tokens.
- ✓Input cost: $0.2 / 1M tokens
- ✓Output cost: $0.2 / 1M tokens
- ✓128k context
DeepSeek R1 Distill Llama 70B 128k (Groq) priced per 1M tokens.
- ✓Input cost: $0.75 / 1M tokens
- ✓Output cost: $0.99 / 1M tokens
- ✓128k context
Qwen3 32B 131k (Groq) priced per 1M tokens.
- ✓Input cost: $0.29 / 1M tokens
- ✓Output cost: $0.59 / 1M tokens
- ✓131k context
Mistral Saba 24B 32k (Groq) priced per 1M tokens.
- ✓Input cost: $0.79 / 1M tokens
- ✓Output cost: $0.79 / 1M tokens
- ✓32k context
Llama 3.3 70B Versatile 128k (Groq) priced per 1M tokens.
- ✓Input cost: $0.59 / 1M tokens
- ✓Output cost: $0.79 / 1M tokens
Llama 3.1 8B Instant 128k (Groq) priced per 1M tokens.
- ✓Input cost: $0.05 / 1M tokens
- ✓Output cost: $0.08 / 1M tokens
- ✓128k context
Llama 3 70B 8k (Groq) priced per 1M tokens.
- ✓Input cost: $0.59 / 1M tokens
- ✓Output cost: $0.79 / 1M tokens
- ✓8k context
Llama 3 8B 8k (Groq) priced per 1M tokens.
- ✓Input cost: $0.05 / 1M tokens
- ✓Output cost: $0.08 / 1M tokens
- ✓8k context
Gemma 2 9B 8k (Groq) priced per 1M tokens.
- ✓Input cost: $0.2 / 1M tokens
- ✓Output cost: $0.2 / 1M tokens
- ✓8k context
Llama Guard 3 8B 8k (Groq) priced per 1M tokens.
- ✓Input cost: $0.2 / 1M tokens
- ✓Output cost: $0.2 / 1M tokens
- ✓8k context
Llama-3.3-70B-Instruct (Tenstorrent) priced per 1M tokens.
- ✓Input cost: $0.4 / 1M tokens
- ✓Output cost: $0.4 / 1M tokens
OpenAI speech-to-text Whisper-large-v3 charged per minute.
- ✓Provider: OpenAI
- ✓Type: STT
- ✓Cost: $0.0025 per minute
Embedding model BAAI/bge-large-en-v1.5 priced per request.
- ✓Provider: Inworld
- ✓Type: Embedding
- ✓Cost: $0.0023
sentence-transformers/paraphrase-multilingual-mpnet-base-v2 embedding priced per request.
- ✓Provider: Inworld
- ✓Type: Embedding
- ✓Cost: $0.0007
On-premises deployment option for Inworld TTS models; contact for pricing.
- ✓Available for: Inworld-TTS-1 and Inworld-TTS-1-Max
- ✓On-prem deployment
- ✓Contact Inworld for pricing
Pros & Cons
✓ Pros
- ✓Very low TTS cost: $5 / 1M characters starting tier
- ✓Sub-250 ms latency and real-time streaming for conversational use
- ✓Instant (zero-shot) voice cloning from 2–15s of audio
- ✓Multilingual support and high quality (reported top ranking in Hugging Face TTS Arena)
- ✓Enterprise options: on-prem and hosted deployments, SOC2/GDPR compliance
- ✓Integrated safety, memory, and knowledge modules included with platform
- ✓Open-source training framework and active research publications
✗ Cons
- ✗Many advanced or on-prem features require contacting sales for pricing
- ✗Some capabilities are marked Preview/experimental (audio markup, model previews)
- ✗Pricing model is usage-heavy and has many per-model/per-token variants (can be complex)
Compare with Alternatives
Related Articles (3)
Inworld AI unveils an equity-free consumer accelerator to help startups scale with mentorship and resources.
Equity-free, six-week accelerator for AI-native startups with mentorship, investor access, and Inworld credits to scale growth.
Equity-free 6-week accelerator for consumer AI startups with $500K+ credits and access to 100+ investors.
