Trace and evaluate LLM executions with Phoenix observability.
npx skills add https://github.com/ricable/mcai --skill phoenix-observability-ricable
Please help me install this Agent Skill. Skill: phoenix-observability Source: https://github.com/ricable/mcai/tree/main/.agents/skills/phoenix-observability Command: npx skills add https://github.com/ricable/mcai --skill phoenix-observability-ricable
This Skill helps teams understand, debug, and improve LLM applications by combining tracing, evaluation, dataset management, experiments, and production monitoring in one open-source workflow.
Use the Phoenix observability skill to trace an LLM application, evaluate its outputs, and review the results in a self-hosted or notebook-based Phoenix instance.
High-intent search queries and answers about installing and using this skill.
You can trace LLM application executions by instrumenting OpenAI, LangChain, LlamaIndex, or Anthropic workflows to capture spans, attributes, and project-level behavior for debugging execution data.
You can evaluate LLM outputs for hallucination and toxicity by running LLM-as-judge evaluations on versioned datasets or existing traces to measure relevance and custom quality criteria.
Yes, OpenTelemetry instrumentation works with LLM observability workflows to capture traceable execution data, enabling project-scoped trace querying across multiple LLM frameworks.
Yes, you can self-host the Phoenix LLM observability instance using either PostgreSQL or SQLite as the backend database for storing traces and evaluation results.
The best way to compare prompts and models is by running experiments that log results back for ongoing production monitoring, allowing you to evaluate different configurations side by side.
trueraInstrument LLM apps to capture traces for evaluation and debugging.
informatico-madridTrace, evaluate, and monitor LLM applications with OpenTelemetry.
mqzkimBuild and execute LLM evaluation pipelines with LLM-as-Judge and customizable metrics.
handsomelong922Collect traces and evaluate LLM outputs with OpenTelemetry instrumentation.
