Integrations for LLM observability
Coralogix AI Observability integrations give you deep insight into your AI applications — monitor, analyze, and optimize performance across LLM providers, frameworks, and code agents. Get end-to-end visibility into AI workloads with proactive issue detection and efficient performance tuning.
LLM providers and frameworks
Coralogix recommends instrumenting your AI applications with OpenTelemetry. The legacy llm-tracekit SDK integrations remain available for now but will be deprecated shortly — migrate to OpenTelemetry.
OpenTelemetry (recommended)
Instrument with the OpenTelemetry GenAI semantic conventions. AI Center accepts spans from any instrumentation that emits the conventions — official OpenTelemetry contrib instrumentations, community libraries like OpenLLMetry, or your own manual instrumentation. See OpenTelemetry integration for AI Center for the supported providers, frameworks, code examples, and span attribute inventory.
Legacy SDK integrations
Deprecation notice
The llm-tracekit SDK integrations below will be deprecated shortly. New projects should use the OpenTelemetry integration instead. Existing projects should plan to migrate — the new approach uses the standard OpenTelemetry open conventions and removes the dependency on Coralogix-specific instrumentation.
- Amazon Bedrock
- Anthropic
- Gemini
- Google ADK
- LangChain
- LangGraph
- LiteLLM
- Mastra
- Microsoft Foundry
- OpenAI
- OpenAI Agents SDK
- Strands Agents
Code agents
- Claude — hub for Claude Code and Claude Cowork
- Claude Code
- Claude Cowork
- Codex CLI
- Copilot CLI
- Cursor
- Gemini CLI
- OpenClaw
AI discovery
Next steps
Start monitoring your AI applications with Monitor AI applications.