Setup Tracing
Get started sending traces to Phoenix from your application.Setup Tracing (Python)
Configure tracing for Python applications using Phoenix OTEL
Setup Tracing (TypeScript)
Configure tracing for TypeScript/Node.js applications
Auto Instrumentation
Automatically capture traces from popular frameworks and LLM providers
Setup Projects
Organize traces into projects for different applications
Setup Sessions
Group related traces into sessions for conversation tracking
Customize Traces & Spans
Add metadata, tags, and custom attributes to your traces for better organization and filtering.Add Metadata, Tags & Users
Set custom attributes on spans created by auto-instrumentors
Prompt Templates
Track prompt templates and variables in your traces
Manual Instrumentation
Create and customize spans for full control over your tracing.Using Tracing Helpers
Use decorators and helpers to easily instrument specific methods in Python
Querying Spans
Extract span data for analysis, evaluation, or export.Extract Data from Spans
Query spans to construct DataFrames for evaluation and analysis
Run a Query
Specify a Project
Query for Documents
Apply Filters
Extract Attributes
Use Pre-defined Queries
Annotate Traces
Add feedback, labels, and evaluation results to your traces.Annotating in the UI
Manually annotate traces through the Phoenix interface
Annotating via Code
Programmatically log annotations using the Phoenix client
Log Evaluation Results
Attach LLM evaluation scores to spans and documents
Evaluating Traces
Run evaluations on trace data
Import & Export Traces
Save, load, and transfer trace data.Export Traces
Export trace data for analysis or backup
Import Traces
Load existing traces into Phoenix
Export Annotated Spans
Export spans with their annotations
Cost Tracking
Monitor token usage and costs for your LLM applications.Cost Tracking
Track token-based costs across your traces