​
Prepare your extraction task

Before we can extract some data, you’ll need to tell the wizard what data you want to extract and how to extract it.
1

Create an extractor for your extraction task

You can just describe the shape of data you want to extract, and an AI will generate an initial draft for you. Create extractor

2

Refine your JSON Schema and output instructions

Edit the generated schema to your liking and add other instructions for the AI to follow. Read more in the Extractors section.Define JSON Schema

Extractors

Extractors are the core configuration objects in Data Wizard


3

Select the LLM to use

You can select from a large number of LLMs thanks to the LLM Magic PHP package. You will need to add your API keys in the LLM settings before you can use them in an extractor. Find out more in the LLM Provider Configuration section.Define JSON Schema

LLM Provider Configuration

Configure your Large Language Model (LLM) API provider in Data Wizard to connect to leading LLMs like OpenAI, Anthropic, Google AI, Mistral AI, and more.


4

Select the extraction strategy to use

There are multiple built-in strategies to choose from, or you can create your own custom strategy.Define JSON Schema

Extraction Strategies

Learn about the built-in and custom extraction strategies available in Data Wizard.


5

Run the extractor

After you have configured your extractor, you can run it to extract data from your documents.You can either use the built-in UI to do this, or you can integrate the feature into an existing application using the iFrame and HTTP API.

Run inside DataWizard

Via Data Wizard’s backend UI

Run inside your own application

Via the embedded iFrame UI

​
Run inside Data Wizard

1

Start an Extraction through a Bucket or an Extractor

All of the data that Data Wizard generates is viewable on the run page. This includes in-progress data as it’s being generated, as well as the final extracted data. You can view the data both as raw JSON or in the GUI derived from the JSON schema.

Launch from Extractor

Launch button in bucket and extractor views

Launch from Bucket
2

View the extracted data in the GUI or as JSON

All of the data that Data Wizard generates is viewable on the run page. This includes in-progress data as it’s being generated, as well as the final extracted data. You can view the data both as raw JSON or in the GUI derived from the JSON schema.
You can use the built-in UI to create and configure your extractor.View Data in GUI
3

Inspect each extraction step

You can inspect each step of the extraction process to see how the AI is interpreting your instructions and what data it is returning.
You can use the built-in UI to create and configure your extractor.View Data in GUI
4

Download the Data as JSON, XML or CSV

asd
5

Restart the extraction with different parameters

asd
6

Modify the data using AI

asd
7

Analyze extraction costs

asd

​
Run inside your own application




Next Steps

Learn how to extract some data

Step by step guide to extract data from documents using Data Wizard.

Extractors

Learn how to define and configure data extraction tasks.

Strategies

Understand different data processing strategies.

LLM Provider Configuration

Set up your Large Language Model API keys.

Integration

Embed Data Wizard into other applications using iFrames or APIs.