HomeToolsMCPHow It WorksStoriesPhilosophyArchitectureStar on GitHub
All Tools
O
MonitoringFreeOpen Source

OPENLIT

OpenTelemetry-native observability for production AI systems

Apache-2.0

ABOUT

Teams shipping LLM and agent applications often have traces, token usage, latency, cost, GPU metrics, and evaluation results scattered across separate tools or missing entirely. OpenLIT instruments AI workloads with OpenTelemetry so teams can debug failures, understand performance, and monitor production behavior end to end without stitching together custom observability plumbing.

INSTALL
pip install openlit

INTEGRATION GUIDE

1. Trace multi-step agent and RAG workflows from user request to final response 2. Monitor token usage, latency, and cost across multiple LLM providers in production 3. Track GPU and vector database health alongside application-level AI traces 4. Run evaluations and compare prompt or model changes before broad rollout 5. Export AI telemetry to existing observability stacks that support OpenTelemetry

TAGS

observabilityllm-observabilitytracingopentelemetryevaluationsgpu-monitoringagents