Langtrace

Langtrace is an open-source observability tool for developers who need to debug LLM applications and track their performance. You can run it locally or on your own servers with Docker and Docker Compose. Its traces follow OpenTelemetry standards.

It captures LLM API calls, framework activity and vector database operations, so you can inspect the steps within an application workflow. Latency, cost and usage metrics help identify slow calls and understand resource use. It also provides evaluations and visual analytics for examining application behavior.

Integration coverage depends on the language. Both the Python and TypeScript SDKs support providers including OpenAI, Anthropic, Gemini and AWS Bedrock, along with LlamaIndex and databases such as ChromaDB, QDrant and PGVector. Python also supports Ollama for local LLM applications, plus Langchain, Langgraph, CrewAI and LiteLLM. Vercel AI support is available through TypeScript.

With self-hosting, Langtrace keeps its observability data on your servers and doesn't collect telemetry. The repository also documents a managed cloud setup using account-based hosted projects. Your application's choice of model provider remains separate from where you host Langtrace.

The application uses the AGPL-3.0 license; its SDKs use Apache 2.0. The self-hosted stack pairs a Next.js application with Postgres for metadata and ClickHouse for spans, metrics, logs and traces.