Documentation

Overview

The API Discovery feature lets you explore and test data source APIs through the AI Agent before implementing report types for extraction. It provides a fast & straightforward way to:

• Explore available API endpoints for supported data sources
• Test API calls with your existing data source credentials
• View API documentation and examples

How to use the API Discovery feature

Prerequisites

To use the API Discovery feature, you'll need:

• Active connection credentials for the data source you want to explore
• Access to the AI Agent

Guide

1. Open the AI Agent chat interface and type your request about exploring a specific API (e.g., "I want to explore Adlabels for Facebook Ads").
2. When prompted, select the API version you want to explore (if multiple versions are available).
3. Select one of your existing connections when prompted to use its credentials in API requests.
4. The AI Agent reviews the API documentation and examples to identify available endpoints and their parameters, then sends requests to the API using the selected credentials.
5. View the API response and continue exploring other endpoints if needed.
6. Each request produces a document automatically created by the AI Agent. This document captures all relevant parameters, headers, and configuration details used during the request.
   1. It can be used later for historical reference, auditing, or troubleshooting purposes.

The AI Agent will guide you through the process, offering relevant documentation and executing test API calls based on your inputs.

Use cases

Direct API Discovery

A marketing analyst wants to explore available API data to understand what campaign information can be retrieved.

Scenario:

The analyst is starting from scratch - no sample data is available. They want to browse and test the API to learn what fields and entities are accessible, such as campaign settings, spend, or targeting.

Goal:

1. Discover available endpoints and data models
2. Explore fields and relationships between entities
3. Preview sample data from each endpoint
4. Identify useful fields for reporting or analysis

Data sample (CSV, JSON, screenshot) to API endpoint mapping

A marketing analyst has a data sample (e.g., CSV or UI screenshot) and wants to find which API endpoints can reproduce it.

Scenario:

The analyst has reporting data from a platform like Reddit Ads, joined across entities (e.g., Campaign, AdGroup), but doesn't know which API endpoints were used. They want to rebuild the same dataset via the API.

Goal:

Automatically analyze the sample data to:

1. Identify relevant API endpoints for each field
2. Detect necessary joins between entities
3. Recommend a sequence of API calls to reconstruct the dataset

Data discrepancy troubleshooting

A marketing analyst notices a mismatch between platform UI metrics (e.g., Google Ads, Facebook) and API-extracted data.

Scenario:

The analyst compares numbers shown in the ad platform's UI with those stored in their warehouse (e.g., BigQuery, S3) and finds inconsistencies. The discrepancy may stem from issues like granularity differences, missing dimensions, unsupported campaign types, or filter mismatches.

Goal:

1. Pinpoint the root cause of data mismatches
2. Compare UI and API data structures and filters
3. Identify unsupported or excluded data (e.g., campaign types)
4. Validate that extraction logic aligns with what’s shown in the UI

Technical details

The API Discovery feature utilizes several key tools and components:

• Documentation vector search: Handles API documentation retrieval and versioning.
• Credentials manager: Provides a list of available credentials for active connections for a specified data source.
• API requests sender: Handles actual API requests and responses.