Run the evaluation (generate with gpt-4.1 then grade)

Lunary API Reference

Getting Started

Github

Langchain

LangChain Integration

Ollama

Ollama Integration

LiteLLM

LiteLLM Integration

IBM WatsonX

IBM WatsonX Integration

Flowise

Flowise Integration

Custom

Step-by-Step Guide : Sending Events via the Lunary API

Observability

Conversations

Chats & Threads

Prompt

Prompt Templates

Classification

Tagging

Users

User Tracking

Feedback

Feedback Tracking

Concepts

Introduction

Retrieve a list of audit logs for the current organization. This endpoint requires that the user has the proper access rights.

Retrieve audit logs

List datasets

Create a new dataset

Get dataset by ID or slug

Delete a dataset

Update a dataset

Create a new prompt

Get prompt by ID

Delete a prompt

Update a prompt

Get prompt variation by ID

Delete a prompt variation

Update a prompt variation

Create a new prompt variation

List evaluations

Create a new evaluation

Get evaluation by ID

Delete an evaluation

Update an evaluation

Create a criterion

Get criterion by ID

Delete a criterion

Update a criterion

List results for an evaluation

Get a single result

This endpoint retrieves a list of users tracked within the project.
It supports pagination, filtering, and sorting options.


List project users

Get a specific user

Delete a specific user

List models

Create a new model

Delete a model

Update a model

The Runs API endpoint allows you to retrieve data about specific runs from your Lunary application.

The most common run types are 'llm', 'agent', 'chain', 'tool', 'thread' and 'chat'.

It supports various filters to narrow down the results according to your needs.

This endpoint supports GET requests and expects query parameters for filtering the results.


Get runs

Get run usage statistics

This endpoint requires a valid private API key sent as a bearer token.


Export runs data

Retrieve detailed information about a specific run by its ID.

Get a specific run

Delete a specific run by its ID. This action is irreversible.

Delete a run

Update run visibility

Update run tags

Update run feedback

Update run score

Get related runs

This endpoint is for reporting data from platforms not supported by our SDKs.

You can use either your project's Public or Private Key as the Bearer token in the Authorization header.

The expected body is an array of Event objects.

For LLM calls, you would first track a `start` event with the `input` data.
Once your LLM call succeeds, you would need to send an `end` event to the API endpoint with the `output` data from the LLM call.

For a full step-by-step guide on sending LLM data to the Lunary API, see the [Custom Integration](/docs/integrations/custom) guide.


Ingest run events

This is the most common endpoint you'll use when working with prompt templates.

This route is used by the Lunary SDKs to fetch the latest version of a template before making an LLM call.

This route differs from all the next ones in that:
- it requires only the `slug` parameter to reference a template
- it doesn't require using a Private Key to authenticate the request (Public Key is enough)


Get the latest version

Update a template version

List all the prompt templates in your project, along with their versions.
Useful for usecases where you might want to pre-load all the templates in your application.


List all templates

Creates a new template with the provided details.
The template includes a slug, mode, content, and additional configuration options.


Create a new template

Get a specific prompt template and all its versions by its ID.


Get a specific template

Delete a template

This endpoint allows you to update the slug and mode of an existing template.
The mode can be either "text" or "openai" (array of chat messages).


Update a template

This endpoint allows you to push a new version of a prompt.
You can specify the content, extra parameters, test values, draft status, and notes for the new version.


Create a new version

Retrieves a list of all views for the current project, ordered by most recently updated.

Getting Started

Audit Logs

Datasets

Evals

API Reference

Users

Models

Runs

Templates

Views

Run the evaluation (generate with gpt-4.1 then grade)

Path Parameters

Response