What dimensions can I use for comparison?

You can compare scores across releases, environments, or specific trace names to see how performance varies across different segments of your application.

Can I detect quality drops between versions?

Yes, the regression detection feature allows you to compare current scores against a baseline period to identify performance dips automatically.

Does it support different time scales?

Yes, you can view trends with hourly, daily, weekly, or monthly granularity depending on your reporting needs.

How do I connect my Langfuse project to this skill?

You must configure the LANGFUSE_PUBLIC_KEY and LANGFUSE_SECRET_KEY environment variables in your terminal or Claude Code configuration.

What is Langfuse Score Analytics used for?

It is used to analyze performance metrics (scores) from Langfuse, helping you track LLM quality, detect regressions, and compare performance across different software releases.

Langfuse Score Analytics

Name: Langfuse Score Analytics
Author: mberto10

bymberto10

0•

Analytics & Monitoring

Analyzes LLM performance metrics and score trends directly within Claude Code using Langfuse observability data.

Langfuse Score Analytics provides deep insights into your AI model's performance by processing and visualizing score data from your Langfuse project. It enables developers to track quality metrics over time, detect performance regressions between software releases, and understand score distributions across different environments or trace types. By integrating these analytics into the Claude Code workflow, teams can make data-driven decisions about model changes and ensure that LLM improvements are measurable and sustained.

Key Features

01Distribution mapping to visualize score spread and identify performance outliers

02Cross-dimensional comparisons across releases, environments, and trace names

030 GitHub stars

04Time-series trend analysis with hourly, daily, weekly, or monthly granularity

05Statistical summaries including mean, percentiles (p50, p95), and min/max values

06Automated regression detection to compare performance against historical baselines

Use Cases

01Generating automated weekly LLM quality reports to monitor accuracy and helpfulness

02Validating new model versions by detecting regressions against previous release baselines

03Comparing performance metrics between development, staging, and production environments

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add mberto10/mberto-compound score-analytics

For use in Claude.ai and ChatGPT

Download Skill