audit_submissions
Categorise a form's recent junk submissions, with the evidence for each.
Reach for this when a form's numbers look good and the leads don't.
It reads a recent window — 200 submissions by default, 1000 at most — and sorts them into bot spam, disposable addresses, duplicates, test junk, and low-quality leads. Every flag carries the evidence that caused it, so you can disagree with it.
It is not a lifetime count. HubSpot serves submissions 50 at a time, so reading all history on a busy form would cost hundreds of requests. Lifetime totals and conversion rates come from form analytics, which needs Marketing Hub Professional; this works on every tier.
Clean submissions come back masked (a***@acme.com) — your legitimate leads' details don't need
to leave the portal. Flagged ones come back in full, because "gibberish name" is not a claim you
can check without seeing the string.
The report ends with a judgement queue: submissions that tripped a signal but no decisive rule. The engine deliberately doesn't guess on those — your AI client rules on them.
Audit a form's recent submissions for junk
Categorise a form's recent submissions: bot spam, disposable addresses, duplicates, test junk, and low-quality leads — each with the evidence that flagged it. Deterministic: the same submissions always grade the same way. Reads a recent window (default 200, max 1000), not lifetime history, and says so in the report. Clean submissions come back masked; flagged ones come back in full so you can see why. The report ends with a judgement queue: submissions with signals but no decisive rule — as the reviewing model, rule on those yourself.
Parameters
| Name | Type | Description | |
|---|---|---|---|
formId | string | required | — |
limit | integer | optional | How many recent submissions to read. Default 200. Each 50 costs one HubSpot request. |