Create User Documentation for Evaluation Features

[maxtechera]

Objective

Create comprehensive user-facing documentation for Datasets, Evaluators, and Evaluations features.

Tasks

Conceptual Overview

• Write "What are Datasets, Evaluators, and Evaluations?" section
• Explain use cases and benefits
• Document when to use Flowise evaluations vs Langfuse
• Add diagrams/visuals if helpful

Step-by-Step Guides

• Create "Getting Started" guide with basic workflow
• Document dataset creation (manual and CSV import)
• Document evaluator setup (simple and LLM)
• Document running evaluations
• Document interpreting results

Reference Documentation

• List available simple evaluator types
• Document LLM evaluator configuration options
• Document evaluation metrics and what they mean
• Document versioning behavior

Best Practices

• Dataset design best practices
• Choosing the right evaluators
• Interpreting evaluation results
• When to re-run evaluations

Location

packages/docs/docs/sidekick-studio/evaluations/ (create new directory)

File Structure (Proposed)

packages/docs/docs/sidekick-studio/evaluations/
├── README.md (overview + getting started)
├── datasets.md (dataset management)
├── evaluators.md (evaluator setup)
├── running-evaluations.md (execution workflow)
├── interpreting-results.md (understanding metrics)
└── best-practices.md (tips and recommendations)

Acceptance Criteria

• Documentation covers all major features
• Includes screenshots or GIFs where helpful
• Step-by-step guides are clear and actionable
• Best practices section provides concrete recommendations
• Documentation integrated into docs site navigation

[linear]

AGENT-604