Manual

Datasets

Datasets are the lists Dregs checks during scoring: disposable email domains, free email domains, suspicious domains, and suspicious IP addresses. Dregs maintains each list, and your team can add its own entries or exclude values that Dregs is treating wrongly in your application.

What Are Datasets

A dataset is a lookup list that analyzers consult during scoring. When an analyzer evaluates an identity, it checks whether a value (an email domain or an IP address) is on a list and factors that into the score. The disposable email check, for example, lowers the Authenticity score when a user's email domain is on the disposable domains list.

Each list has two layers. The global layer is maintained by Dregs and shared by every customer. Your layer is private to your team and sits on top of it, so the entries you add change how scoring treats your users without affecting anyone else.

Built-In Datasets

These are the datasets your team can view and adjust:

Dataset Slug What it holds Matching
Disposable Email Domains disposable-email-domains Disposable or temporary email domains Case-insensitive
Free Email Domains free-email-domains Free or public email providers Case-insensitive
Suspicious Domains suspicious-domains Domains flagged as suspicious or high-risk Case-insensitive
Suspicious IPs suspicious-ips IP addresses flagged as suspicious Case-sensitive

The disposable and free email checks also include lists Dregs curates internally, so a domain can be treated as disposable even when it doesn't appear in the public list you see. An exclusion on your layer still wins, as described below.

Case-insensitive datasets match Example.com, example.com, and EXAMPLE.COM as the same entry. The matching mode is fixed for each dataset.

Including and Excluding Values

Every entry on your layer either includes a value or excludes it:

  • Include adds a value to the list for your team. Use it for a disposable domain Dregs hasn't caught yet, or an IP address from your own investigations.
  • Exclude tells Dregs to stop treating a value as listed for your team. Use it for false positives, such as a regional email provider that a public list marks as disposable but your real customers use.

When your layer has an entry for a value, it decides the result, whatever the global layer says. This works both ways: you can include a value the global list doesn't have, or exclude one it does.

Managing Datasets in the Dashboard

Open Settings and choose the Datasets tab. Each dataset shows how many entries your team has added. Admins can click Manage to add a value with Include or Exclude, and to remove an entry. Other team members can see the lists but not change them.

Adding the first entry to a dataset creates your team's layer for it, and that counts toward your plan's custom dataset allowance. Custom datasets come with the Advanced plan. See Pricing.

Managing Datasets via the API

The REST API exposes the same operations, one entry at a time. Each call takes a scope query parameter. Use CUSTOMER, which covers your team's layer.

  • GET /api/datasets?scope=CUSTOMER lists the datasets with your entry counts.
  • GET /api/datasets/{slug}/entries?scope=CUSTOMER pages through your entries (200 per page by default).
  • POST /api/datasets/{slug}/entries?scope=CUSTOMER with {"key": "example.com", "value": "true"} adds or replaces an entry. A value of "false" makes it an exclusion. Requires an admin or a secret key.
  • DELETE /api/datasets/{slug}/entries?scope=CUSTOMER&key=example.com removes an entry. Requires an admin or a secret key.

There is no bulk import endpoint. To load a long list, call the add endpoint once per value.

Dataset entries feed the analyzers behind the four scores. See Scoring for how analyzers produce observations. Contact dregs@dregs.com to learn more about enabling custom analyzers.