Langfuse is an open-source LLM engineering platform for debugging, monitoring, and improving LLM applications. Trace, evaluate, and manage your AI applications with comprehensive observability.
Triggers when a new trace is created in Langfuse. Polls the traces list ordered by newest first.
2 parameters
Trace Name Filter
Environment Filter
New Score
Triggers when a new score is created in Langfuse. Useful for reacting to quality evaluations.
1 parameters
Score Name Filter
Actions available for Langfuse on Stepper
Ingest Trace
Creates a new trace in Langfuse. A trace represents a single request or operation in your LLM application.
12 parameters
Trace Name
Trace ID
User ID
Session ID
Input
Output
Metadata
Tags
Environment
Version
Release
Make Public
Log Generation
Logs an LLM generation (model call) as an observation within a trace. Captures model, tokens, cost, and input/output.
17 parameters
Generation Name
Generation ID
Trace ID
Model
Input
Output
Prompt Tokens
Completion Tokens
Total Tokens
Start Time
End Time
Metadata
Environment
Version
Level
Prompt Name
Prompt Version
Create Score
Creates a score (quality signal) for a trace or observation. Use to record evaluation results, user feedback, or automated quality metrics.
8 parameters
Score Name
Score Value
Trace ID
Session ID
Observation ID
Data Type
Comment
Environment
Create Prompt
Creates a new version of a prompt template in Langfuse. If the prompt name already exists, a new version is added. Supports both text and chat prompt formats.
5 parameters
Prompt Name
Prompt Type
Labels
Tags
Commit Message
Get Trace
Retrieves a trace by its ID, including all associated observations.
1 parameters
Trace ID
List Traces
Retrieves a paginated list of traces with optional filters.
9 parameters
Trace Name
User ID
Session ID
Environment
From Timestamp
To Timestamp
Tags
Version
Release
Make HTTP Request
Make an HTTP request to any URL with full control over method, headers, and body.