A FlexCard is a face. An OmniScript is a conversation. But neither can do anything without data: records to show, fields to fill, answers to save. Something has to fetch that data from Salesforce and put it back when the work is done. That something is the DataRaptor, the quiet data layer beneath everything else in OmniStudio. It is less glamorous than the parts users see, but learn it well and the whole toolset starts working for you.
What a DataRaptor does
A DataRaptor reads, reshapes, and writes Salesforce data. It sits between your screens and your database, translating between the JSON that FlexCards and OmniScripts speak and the records that live in Salesforce objects.
There are four kinds, and the names tell you the job:
- Extract reads data out of Salesforce.
- Transform reshapes data without necessarily reading or writing records.
- Load writes data into Salesforce.
- Turbo Extract is a faster, streamlined read for simple, single-object lookups.
Most beginners spend their first weeks with Extract and Load, the read and the write. The other two arrive naturally once you need them.
Extract: reading data
An Extract DataRaptor pulls records and hands them back as JSON. You tell it which object to query, which fields to return, and which filters to apply, all through configuration. Then you map each Salesforce field to an output name your FlexCard or OmniScript will use.
Say you extract an account. The output might look like this:
{
"Account": {
"Name": "Coastal Logistics",
"Status": "Active",
"Phone": "555-0142"
}
}Your FlexCard then merges those values with tokens like {Account:Name}, and your card shows real data. This output JSON is the contract between the data layer and the display layer. When a card shows up blank, the mismatch is almost always between the names here and the tokens on the card.
A DataRaptor does not just move data. It shapes it into exactly the form your screen expects, so the screen can stay simple.
Load: writing data
A Load DataRaptor does the reverse. It takes a bundle of JSON, often the data JSON an OmniScript has collected, and writes it into Salesforce records. You map each incoming field to the object and field it belongs to, and the DataRaptor handles the insert or update.
This is the satisfying end of a guided flow. The user finishes the last step, presses submit, and a Load DataRaptor quietly saves their work. No Apex required, just a careful mapping.
Transform and Turbo Extract
Transform is for reshaping. Sometimes data arrives in one structure and your screen needs another, perhaps flattening a nested response or renaming fields. A Transform DataRaptor restructures the JSON without touching the database.
Turbo Extract is a lighter, faster Extract. When you only need fields from a single object with straightforward filters, Turbo Extract skips some of the heavier processing and returns data quickly. Reach for it when the read is simple and speed matters.
How DataRaptors feed everything else
Here is the picture worth keeping. A FlexCard calls an Extract to get the data it displays. An OmniScript calls an Extract at the start to pre-fill fields and a Load at the end to save the result. When a job needs several reads and writes coordinated together, an Integration Procedure calls multiple DataRaptors in sequence on the server.
In every case, the DataRaptor is the hands that reach into Salesforce. The other tools decide what to do; the DataRaptor does the touching.
A note for those coming from SOQL
If you have written queries before, a DataRaptor Extract will feel familiar in spirit. You are choosing an object, picking fields, and filtering rows, which is the same thinking behind a SELECT statement. If that comparison helps you, SOQL for Beginners is a good grounding read. The difference is that a DataRaptor wraps that querying in configuration and hands you JSON instead of a result set, ready for a card or a script to consume.
Start small. Build one Extract, point it at one object, and look at the JSON it returns. Then build one Load and watch a record update. Once those two clicks feel reliable, the data layer of OmniStudio holds no more mysteries for you.
Your next step
See how the data layer connects upward and outward: