Documentation Index
Fetch the complete documentation index at: /llms.txt
Use this file to discover all available pages before exploring further.

Skip to main content

Phoenix home page

SDK & API Reference

Cookbooks

Agent Workflow Patterns

Overview
AutoGen
CrewAI
Google GenAI SDK (manual orchestration)
OpenAI Agents
LangGraph
Smolagents

AI Engineering Workflows

Iterative Evaluation & Experimentation Workflow (Python)
Iterative Evaluation & Experimentation Workflow (TypeScript)
Aligning LLM Evals with Human Feedback (TypeScript)
Prompt Optimization Techniques
Run Repetition Experiments

Tracing

Identifying High-Signal Traces in Production
Agentic RAG Tracing
Generating Synthetic Datasets for LLM Evaluators & Agents
Structured Data Extraction
Product Recommendation Agent: Google Agent Engine & LangGraph
OpenInference Best Practices
More Cookbooks

Human-in-the-Loop Workflows (Annotations)

Using Human Annotations for Eval Driven Development
Aligning LLM Evals with Human Feedback (TypeScript)

Prompts

Optimizing Coding Agent Prompts - Prompt Learning
Optimizing Prompts for LLM Classification - Prompt Learning
Few Shot Prompting
ReAct Prompting
Chain of Thought Prompting
Prompt Optimization Techniques
LLM-as-a-Judge Prompt Optimization

Evaluation

Write Your First Custom Eval
OpenAI Agents SDK Cookbook
Evaluate a Talk-to-Your-Data Agent
Trace-level Evaluation: Beyond Input/Output Checks
Session-level Evaluation: Scoring the Whole Conversation
Evaluate RAG
Code Readability Evaluation
Relevance Classification Evaluation
Using Ragas to Evaluate a Math Problem-Solving Agent
More Cookbooks

Guardrails & Safety

Jailbreak and Prompt Injection Defense
Designing Realtime Guardrails: Input, Output, and the Cost of Blocking

Datasets & Experiments

Analyzing Customer Review Evals with Repetition Experiments
Experiment with a Customer Support Agent
Comparing LlamaIndex Query Engines with a Pairwise Evaluator
Prompt Template Iteration for a Summarization Service
Text2SQL Experiments
Why Public Benchmarks Lie: Building Your Own Eval Harness
More Cookbooks

Use Cases

Datasets & Experiments

More Cookbooks

Iteratively improve your LLM task by building datasets, running experiments, and evaluating performance using code and LLM-as-a-Judge.

Use Cases

Answer and Context Relevancy Evals
RAG with Reranker
Response Guideline Evals

Was this page helpful?

Suggest edits Raise issue

Why Public Benchmarks Lie: Building Your Own Eval Harness

Phoenix home page

slack x bluesky linkedin github

English 日本語 (Japanese)中文 (Chinese)

slack x bluesky linkedin github

slack x bluesky linkedin github

Assistant

Responses are generated using AI and may contain mistakes.