What is the Langfuse CLI skill?

It is a specialized capability for Claude that provides a command-line interface to the Langfuse platform, allowing users to monitor, trace, and evaluate LLM applications.

How does it help with prompt management?

The skill allows you to create, list, and label prompt versions, making it easy to manage prompt iterations and deploy specific versions to production without code changes.

Does it support model evaluation?

Yes, you can create and manage evaluation datasets, add items from existing traces, and view dataset runs to benchmark model performance.

Can I export trace data to other tools?

Yes, the skill supports exporting data in multiple formats including table, JSON, CSV, and Markdown for external analysis or documentation.

Can I use this to track LLM costs?

Yes, you can query aggregated metrics for total cost, input/output tokens, and cost-per-model using integrated commands like 'lf metrics query'.

Langfuse CLI

Name: Langfuse CLI
Author: tavva

bytavva

Analytics & Monitoring

Monitors LLM observability, manages prompt versioning, and handles evaluation datasets through the Langfuse platform.

About

The Langfuse CLI skill empowers Claude to interact directly with the Langfuse LLM observability platform to monitor application performance and manage production prompts. It enables real-time querying of traces, sessions, and metrics like latency and token costs, while also providing robust tools for prompt management and dataset evaluation. Whether you are debugging specific LLM calls, tracking production costs, or deploying new prompt versions with labels, this skill bridges the gap between local development and production monitoring by bringing observability data directly into your workflow.

Key Features

Analyze LLM costs, token usage, and latency with granular metrics
Export data in multiple formats including JSON, CSV, and Markdown
Manage prompts with version control, labels, and production deployment
0 GitHub stars
Query and filter LLM traces and sessions for deep observability
Create and manage evaluation datasets and runs for model testing

Use Cases

Investigating production errors by retrieving detailed trace observations and scores
Versioning and deploying new prompt templates with labels like 'production' or 'staging'
Tracking and optimizing LLM API expenditures across different models and environments

About

Key Features

Analyze LLM costs, token usage, and latency with granular metrics
Export data in multiple formats including JSON, CSV, and Markdown
Manage prompts with version control, labels, and production deployment
0 GitHub stars
Query and filter LLM traces and sessions for deep observability
Create and manage evaluation datasets and runs for model testing

Use Cases

Investigating production errors by retrieving detailed trace observations and scores
Versioning and deploying new prompt templates with labels like 'production' or 'staging'
Tracking and optimizing LLM API expenditures across different models and environments