Instrument LLM applications with tracing, prompt management, and evaluation using Langfuse.
npx skills add https://github.com/bilacchi/agents-skills --skill langfuse-bilacchi
Please help me install this Agent Skill. Skill: langfuse Source: https://github.com/bilacchi/agents-skills/tree/main/skills/langfuse Command: npx skills add https://github.com/bilacchi/agents-skills --skill langfuse-bilacchi
High-intent search queries and answers about installing and using this skill.
Import openai from langfuse.openai instead of the standard SDK and use it as a drop-in replacement. All chat completions are automatically traced, and you can pass name, session_id, user_id, tags, and metadata as extra parameters.
Create a Langfuse CallbackHandler with your keys and pass it via config={"callbacks": [handler]} when invoking chains, agents, or retrievers. Every LLM call and tool use becomes a nested span in the trace.
Yes, Langfuse offers both a managed cloud and a self-hosted option. Self-hosting requires your own infrastructure, and you point the SDK to your instance by setting the host parameter instead of the cloud URL.
Attach scores to traces manually, or use an LLM-as-judge pattern where a cheaper model rates responses. You can also create datasets with expected outputs, run evaluations per item, and link traces to dataset runs for comparison.
Traces go missing when the process exits before data is sent. Call langfuse.flush() before the function returns to force delivery of queued events, which is essential in serverless and short-lived environments.
ihatesea69Trace LLM calls, evaluate outputs against datasets, and monitor production AI applications with LangSmith observability.
Clay-HHKTrace, evaluate, and monitor LLM applications using the LangSmith Python SDK.
camgrimsecAnalyze code to map attack surfaces and trace data flows from source to sink.
HemantSudarshanInstrument LLM calls for tracing and observability in Python or TypeScript pipelines.
