OpenLLMetry
traceloop.com
- Free plan
- Web
- API
- Self-hosted
What OpenLLMetry is
OpenLLMetry is a free, open-source project for monitoring and debugging LLM application execution. It adds non-intrusive tracing built on OpenTelemetry, with annotations for workflows, tasks, agents, and tools. Traces can be exported to Traceloop or an existing observability stack, including Datadog, Dynatrace, Honeycomb, New Relic, Sentry, Splunk, and OpenTelemetry Collector. SDK guides cover Python, JavaScript/TypeScript, Go, and Ruby. Supported providers include OpenAI, Anthropic, Amazon Bedrock, and Google Gemini, while supported frameworks include LangChain, LlamaIndex, CrewAI, Haystack, and OpenAI Agents; language coverage varies. Prompts, completions, and embeddings are logged by default, with content logging disableable through TRACELOOP_TRACE_CONTENT=false. Multi-modal tracing can capture images, audio, video, documents, and mixed content, though visualization currently requires Traceloop. Media files over 10MB may be truncated or linked. Apache-2.0 licensing permits free use.
Who it suits
OpenLLMetry suits developers who need tracing for LLM, agent, or retrieval workflows and want to send data to Traceloop or another observability backend. It supports self-hosted and hosted deployment options.
Good to know, and what to watch
- OpenTelemetry-based tracing is non-intrusive
- SDK guides cover four programming languages
- Supports major providers and agent frameworks
- Content logging can be disabled
- Apache-2.0 licensed and free to use
- Multi-modal visualization requires Traceloop
- Media over 10MB may be truncated or linked
- Provider and framework language support varies
- Content logging is enabled by default
Plans
Every plan and priceEnterprise
Free Forever
Free plan
The verdict
OpenLLMetry provides broad tracing coverage with an open license and multiple export destinations. Teams handling multi-modal data should note the Traceloop visualization requirement and large-file limitation.
What the maker publishes
- Backend integrations
- The integration catalog includes Datadog, Dynatrace, Honeycomb, New Relic, Sentry, Splunk, and OpenTelemetry Collector. source
- Content logging
- OpenLLMetry logs prompts, completions, and embeddings by default, and content logging can be disabled with TRACELOOP_TRACE_CONTENT=false. source
- Export destinations
- Traces can be exported to Traceloop or an existing observability stack. source
- Framework support
- The supported framework list includes LangChain, LlamaIndex, CrewAI, Haystack, and OpenAI Agents, with language support varying by framework. source
- Languages
- The documentation provides SDK getting-started guides for Python, JavaScript/TypeScript, Go, and Ruby. source
- License
- The pricing page says OpenLLMetry is Apache-2.0 licensed and free to use. source
- Multi-modal size limit
- The documentation says very large media files over 10MB may be truncated or linked rather than embedded. source
- Multi-modal tracing
- Multi-modal tracing can capture images, audio, video, documents, and mixed content; its visualization is currently available only with Traceloop as the observability backend. source
- Provider support
- The supported list includes OpenAI, Anthropic, Amazon Bedrock, Google Gemini, and other LLM providers, with Python and TypeScript support varying by provider. source
- Purpose
- OpenLLMetry is an open source project for monitoring and debugging LLM app execution, with non-intrusive tracing built on OpenTelemetry. source
- Self-hosting
- Traceloop says it supports on-premise deployment in AWS, GCP, Azure, Kubernetes, air-gapped, and other restricted environments. source
- Support
- The Enterprise plan includes dedicated Slack support. source
- Tracing features
- Workflow annotations can describe workflows, tasks, agents, and tools in traces. source
Checked against the maker's pages on 8 Oct 2026.