Overview

Atla

Running your first eval with a few lines of code

Quickstart

Intro to evals

Models

End-to-end tutorials using Selene with the Atla API

Tutorials

Atla Trust Center

Getting the most out of your Atla evaluator

Best Practices

Langfuse

DeepEval

HuggingFace

Ollama

Introduction

MCP Server

Python Library

Contact

Discord

API Reference

Sign Up

Using the Eval Copilot, formerly known as the Alignment Platform

Eval Copilot (beta)

Defining your scoring criteria

Gathering the right test cases for evaluation

Develop your test cases

Unlock scalable oversight by aligning Selene to experts

Align Selene with expert scores

Simple and flexible way to run evaluations

Run your evaluations

Learn valuable insights that you can readily action on

Generate insights

Default Metrics

Custom metrics

Retrieval-Augmented Generation (RAG)

Evaluating against known reference answers

Ground Truth

Run an evaluation directly via the Atla evaluation service.

Create Evaluation

Use OpenAI chat completions endpoint to run evaluations

Creating, managing and using custom evaluation metrics.

Create Metric

Delete Metric

Get Metric

List all metrics for a project. Optionally include Atla default metrics that the
project has access to.

Get Started

Learn about Selene

Build with Selene

Integrations

Self-deployment

Overview