Welcome to Lumina
Production-grade observability for AI applications. Lumina is an OpenTelemetry-native platform for monitoring LLM applications in production. Track costs, latency, and quality across distributed AI systems with full trace visibility and regression testing.Get Started
Quickstart
Install Lumina and send your first trace in under 10 minutes
SDK Reference
Instrument your application with the TypeScript/JavaScript SDK
Core Concepts
Learn fundamental concepts: traces, spans, and attributes
Production Deployment
Deploy to production with Kubernetes
Features
Cost Tracking
Automatic cost calculation for OpenAI, Anthropic, and major providers
Distributed Tracing
Hierarchical tracing for RAG pipelines, agents, and multi-model systems
Quality Monitoring
Detect semantic degradation with built-in similarity scoring
Replay Testing
Test changes against real production data before deployment
Real-Time Alerts
Get notified of cost spikes and quality drops in under 500ms
Self-Hosted
Deploy on your infrastructure with full data control
Popular Pages
Multi-Span Tracing
Track complex workflows like RAG pipelines with hierarchical spans
OpenAI Integration
Complete guide to integrating OpenAI models
Anthropic Integration
Complete guide to integrating Anthropic Claude models
Performance Optimization
Optimize Lumina for high-throughput workloads
Community
GitHub
Star us on GitHub and contribute to the project
Discussions
Ask questions and get help from the community
Report Issues
Report bugs and request features
Examples
Explore example applications
Self-Hosted
The open-source version is free forever with:- 50,000 traces per day — Resets at midnight UTC
- 7-day retention — Automatic cleanup
- All features included — No paywalled functionality
Need Help?
FAQ
Frequently asked questions about Lumina
Troubleshooting
Common issues and solutions