This post is outdated.
We've made big improvements since this was published.
Check out what's new:
Check out what's new:
AI text-to-speech for video game characters
Integrate real-time or pre-recorded AI voices into your games with Inworld's realistic text-to-speech. Inworld's voices have unmatched emotional depth and realism \u2013 and they're more cost-effective than other AI voice generators.
Why Inworld Voice?
Enhance your game with our NPC voice generator
Expressive voices: Sick of robotic AI voices? Choose emotionally resonant voices that capture the personalities of your villains, heroes, or orcs.
Gaming-native: Don’t choose a TTS model made for chatbots. Get voices made for common gaming archetypes like sages, soldiers, bosses, and sorceresses.
Localization: Ensure your game reaches a global audience with multilingual voice support.
Use cases
3 ways to use Inworld voices for games
Voice acting: Choose from our extensive video game voices or request a custom cloned voice to record character dialogue for your game.
Real-time voice API: Use our AI voice API to power real-time text-to-speech in your games.
With our AI Engine: Use our AI Engine to power your game’s AI agents. Our expressive voices are included.
Inworld difference
The best NPC text-to-speech solution
High quality: We created a new benchmark for measuring audio quality and expressiveness across five categories: generation accuracy, prosody, talking speed, expressiveness, and speaker similarity.
Ultra low latency: Inworld’s AI voices boast 250ms end-to-end 50th percentile latency for approximately 6 seconds of audio generation.
Cost-effective: Our pricing is significantly more affordable than other expressive TTS solutions.
Real-time voice API
Don’t let latency wreck your gameplay
Easy integration: Seamlessly integrate our TTS API using easy-to-use REST or gRPC APIs – with either basic or JWT authentication – all backed by extensive documentation.
Scalability and reliability: Inworld’s text-to-speech API is designed to reliably handle high volumes of requests to ensure uninterrupted AI speech synthesis.
Developer-friendly tools: Comprehensive SDKs and documentation to get you up and running fast.
AI voice cloning for games
Get custom voice cloning
Work with voice actors: Want to license the voice of your existing voice actors for your games? We’ll work with you to create a custom voice model.
Custom training = better voices: Get AAA-quality with a custom cloned model. We test and iterate to ensure you get higher quality voices.
Perfect for all studio sizes
Passion projects: Leverage Inworld’s affordable TTS to integrate recorded dialogue or real-time voices into your game.
Indies: Add voices or localize your game to expand your market – and revenue.
AAA: Get AAA-quality voices to power AI NPCs, cloned voices for last-minute line changes, localization support, and more.
Use cases: From NPCs to narration – and more!
Game characters: Give your characters emotionally resonant voices.
Narration: Voice your game narrative with an expressive voice.
In-game announcements: Get gamers’ attention with in-game announcements.
Frequently asked questions
We will be launching Mandarin, Korean, and Japanese voices soon. We will also be rolling out additional voices in the next few months. Stay tuned!
Yes! We will be rolling out self-serve voice cloning in the future. However, for the best results, we recommend custom voice training – which is currently available. Get in touch for more info.
We can train a cloned model with as little as 20 minutes of audio. However, we recommend more for the best results.
You can use AI to voice video game characters via API or the Inworld AI Engine. With the API, you can either record voices during production or use it for real-time text-to-speech in your game. If you use the Inworld AI Engine, which powers all aspects of AI agent behavior, expressive voices are included.
Inworld has the best gaming-native AI voice generator. Our voices are expressive, cost-efficient, and high-quality. Key features include high-quality speech synthesis, gaming-native voice archetypes, ultra-low latency, and multilingual support.
We make the first 100 daily requests free via the text-to-speech API. Reach out to our team for additional pricing options for your scale and application.