Runtime API

One API for conversational AI experiences designed for children. You build the product. We handle speech, safety, and compliance.

Try It Now View Documentation

Focus on your product

Building AI for kids means solving speech recognition, content safety, regulatory compliance, and real-time performance simultaneously. UG handles all of it so you can focus on the experience.

  • Child-native speech recognition (95%+ accuracy, ages 3-16)
  • Real-time content safety and moderation
  • COPPA, GDPR-K, and kidSAFE compliance built in
  • Sub-2 second voice-to-voice latency
  • Cross-platform SDKs (iOS, Android, Unity, Web)
import { ConversationManager } from 'ug-js-sdk'; const conversation = new ConversationManager({ apiKey: 'your-api-key' }); conversation.on('text-event', (e) => console.log(e.text)); await conversation.start(); await conversation.sendText('Hello!');

What the Runtime provides

A complete stack for child AI experiences, designed for production.

Child Speech Recognition

95%+ accuracy on ages 3-16. Handles developmental speech patterns, background noise, and natural conversation.

Real-time Safety

Content filtering, PII detection, and behavioral guardrails enforced at the infrastructure level.

Low Latency

Full voice-to-voice loop in under 2 seconds. ASR, LLM, and TTS optimized for natural conversations.

Cross-Platform SDKs

Native SDKs for iOS, Android, Unity, Unreal, and Web. Same API, optimized for each platform.

Character Studio

Define personalities, behaviors, and safety boundaries. Test and iterate before deploying to production.

Production Ready

99.9% uptime SLA, auto-scaling, global edge deployment, and real-time analytics.

You build the experience.
We handle the infrastructure.

UG Handles

  • Child speech recognition and understanding
  • Real-time content safety and moderation
  • COPPA, GDPR-K, and kidSAFE compliance
  • Voice synthesis and audio streaming
  • PII detection and data protection
  • Parental consent infrastructure
  • Conversation memory and context
  • Scaling, uptime, and global deployment

You Build

  • Your character's personality and voice
  • The user experience and interface
  • Game mechanics or learning objectives
  • Your app's unique features
  • User acquisition and engagement
  • Your business model

Real-time WebSocket API

Stream messages over WebSocket for instant, natural conversations.

Send WSS /interact
{ "kind": "set_configuration", "config": { "ref": "captain_cosmos_character" } } { "text": "Hi! What's your name?" }
Stream Response text events
{ "type": "stream", "event": "text", "text": "Hi there, space cadet!" } { "type": "stream", "event": "text", "text": " I'm Captain Cosmos. Ready to explore the galaxy together?" }

Performance Benchmarks

Measured under production load. January 2026.

Metric Result
Avg Text Response
Time to first token
 <1s
Avg Audio Response
Text input to audio output
 <2s
Avg Audio to Audio
Full voice-to-voice loop
 <3s
Platform Uptime
Rolling 90-day average
 99.999%
PII Stored
Child personal data retained
 0

Start building today

Get started with the documentation and try the API for free.

Try the API View Documentation