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

# AI Extract

> Schema-driven extraction for turning webpages and source content into structured data.

# AI Extract

AI Extract is designed for **structured extraction**.

When you already know what content you want to inspect, AI Extract helps you turn that source into predictable machine-readable output.

## What AI Extract is for

Use AI Extract when you want to:

* extract fields from a webpage
* convert unstructured content into JSON
* normalize profile, company, or product data
* capture information into a predefined schema
* power internal data collection workflows

Common examples:

* extract company information from a website
* parse a profile page into structured fields
* collect product data from landing pages
* transform rich page content into normalized records

## How it works

You provide the target content plus the structure you want back.

NeuralVerge then:

1. reads the source
2. applies the extraction logic
3. maps the result into the requested schema
4. returns structured output you can use in downstream systems

## Why teams use AI Extract

AI Extract is useful when free-form summaries are not enough.

It helps teams build reliable pipelines for:

* CRM enrichment
* lead qualification
* catalog building
* internal databases
* compliance intake workflows

## Inputs

Extraction workflows typically include:

* a source or page to inspect
* extraction instructions
* a JSON schema or structured field definition
* optional settings that control extraction behavior

## Outputs

AI Extract responses can include:

* a `human` explanation or summary
* a `machine` payload containing extracted fields
* schema-aligned JSON for downstream usage
* usage metadata such as points

## When to choose AI Extract

Choose AI Extract when:

* you already know the source to inspect
* you need a defined output shape
* the main goal is structured data capture

If you need broad exploration across many sources, use [AI Research](./ai-research) instead.

If you need a provider-specific lookup like LinkedIn or email enrichment, use [Data Sources](./data-sources).

## Related API endpoints

In the API Reference, AI Extract includes the extraction endpoint used for structured runs.