Skip to content

Pull connectors and visual binder

Connectors are not finished yet

Connectors work end to end, but the configuration vocabulary is still developer-facing and the dry-run flow is being revised. Showcase marks Integrations as coming soon in the editor for that reason. The rest of Audience Integrations is generally available.

A pull connector fetches a configured source on demand or schedule, snapshots the response, parses it, and applies the same mapping engine used by push ingest.

Configure pull

  1. Create a connector and set How data arrives to Showcase fetches data from a URL.
  2. Enter the Source URL and Refresh every (minutes) (from 1 minute up to 7 days). Showcase accepts http:// and https:// sources; use HTTPS wherever the source supports it.
  3. Choose JSON, XML, CSV, or HTML parsing.
  4. Save, then use Run now to verify connectivity.
  5. Inspect last-run status before enabling a schedule.

Under each pull connector's row, the Schedule panel sets how often the connector runs. Enter the number of minutes after Every, then select Save. Run now fetches the source at once. The panel also shows the Next run time and, after a failed run, Last: failed. From Showcase 26.11.1 the panel is readable in the dark theme; before that, it showed near-white text on a near-white row.

The Integrations page in the dark theme: the Connectors table with a pull connector row, and under it the Schedule panel with Every 60 minutes, Save, Run now and Next run: not scheduled

Showcase blocks loopback, private, link-local, and otherwise unsafe destinations. Redirects are checked again. By default a single fetch is limited to 5 MB and 10 seconds. If a legitimate enterprise source is blocked or too large, contact MultiTaction support with the required origin rather than weakening network controls broadly.

Visual binder

  1. Select Visual binder on the pull connector's row in Integrations.
  2. Load a server-fetched snapshot.
  3. Select the repeating item container.
  4. Click a representative field in the sanitised preview.
  5. Bind exactly one display-name field, then external ID and optional properties/media.
  6. Enter the Target root node ID.
  7. Select Dry-run. Review the create, update, archive and quarantined counts.
  8. Select Apply mapping only after the dry-run succeeds.
  9. Reload the page and confirm bindings remain.
  10. Run the connector and verify that the saved mapping, not an unsaved draft, is used.

The snapshot iframe removes scripts and external references. It is not the live source page.

Snapshot tickets

Snapshot preview URLs contain short-lived signed tickets. They are scoped to the connector snapshot and expire automatically. Do not copy them into manuals or messages. A missing, changed, forged, or expired ticket must fail; refresh the snapshot from the binder instead of extending a URL manually.

Troubleshooting

  • Fetch blocked: source resolved to a private/loopback address, redirect escaped policy, DNS failed, or allowlist rejected it.
  • Snapshot empty: confirm content type, authentication available to the connector, and parser record path.
  • Apply mapping disabled: a draft change has not been dry-run, the target root node ID is missing, or required name/external-ID bindings are incomplete.
  • Mapping disappears after reload: select Apply mapping after the final dry-run, then check the connector's run history.