Skip to main content
Inworld Portal provides a comprehensive suite of tools designed to help developers test, optimize, and deploy AI-driven applications with confidence. By combining powerful observability features with intuitive evaluation capabilities, Portal enables teams to build high-quality, production-ready AI experiences efficiently.

Experiments

Run A/B experiments without code deployment using our Experiments system:
  • Variant Management: Upload different graph configurations and test multiple variants simultaneously
  • Targeting Rules: Control which users experience which variants based on user attributes
  • Traffic Distribution: Gradually roll out changes with precise traffic allocation controls
  • Instant Deployment: Switch between models, prompts, and graph configurations in real-time

Dashboards

Monitor your application performance and experiment results with comprehensive dashboards:
  • Default Metrics: Track essential performance indicators like graph execution counts, error rates, and latency percentiles
  • Custom Metrics: Define and visualize application-specific metrics using our telemetry system
  • Real-time Monitoring: Get instant visibility into your application health and experiment performance

Logs & Traces

Monitor and analyze AI interactions in real-time with our advanced logging and tracing systems:
  • Logs: Review historical data to monitor errors and debug issues.
  • Traces: Understand the flow of your application with detailed execution traces. Use them to identify latency bottlenecks and debug issues when they arise.

Prompt & Model Tuning

Refine AI behavior through our interactive testing and prompt iteration tools:
  • LLM Playground: Test language models with various prompts and variables. Compare outputs across different configurations to identify the optimal settings.
  • TTS Playground: Generate AI-powered speech with customizable parameters for speed and emotional tone.
  • Model Store: Browse and access the latest AI models available through Inworld, with detailed information about capabilities, performance characteristics, and recommended use cases.

Next Steps