Instantly's health score is a useful warmup signal. It is not inbox placement, and it is not a blacklist check.
What the score is
On API v2, warmup analytics return landed_inbox, landed_spam, and a health_score for the accounts you ask about. That score describes how Instantly's own warmup network classified recent warmup mail. The daily account analytics endpoint is a different report: sent, bounced, and replies for campaign mail, per mailbox, per day.
Both are worth having. Neither one tells you that the SPF record was deleted this morning, and a warmup score is not a DNSBL lookup. Inbox Watch does not query Spamhaus. The result page links to their checker.
What still fails while the score looks fine
- The domain is listed on NordSpam, or a published mail IP is on PSBL or a Mailspike block code.
- DKIM was rotated or removed, so new campaign mail is unsigned.
- DMARC was relaxed to
p=none. - Campaign bounce rate jumped because the list rotted, while warmup mail still lands.
- Human replies dropped because prospects never saw the mail. Warmup replies are not prospect replies.
People lose weeks here. The dashboard they stare at is measuring a different pile of mail than the campaigns that pay the bills.
How Inbox Watch uses Instantly
The adapter calls two official v2 endpoints and nothing else:
GET /api/v2/accountsto list mailboxes. The key needsaccounts:read.GET /api/v2/accounts/analytics/dailyfor sent, bounced, and unique replies. The date window stays inside Instantly's 31-day maximum, and email filters stay at or below 100 addresses per request.
Reply rate in Inbox Watch is unique replies divided by emails sent, for the monitored mailboxes on that domain. The baseline is the first earlier window that contains at least 80 sent emails, excluding the most recent 7 days. It does not slide down when the domain gets worse, or the alert would never fire.
The hourly job then re-checks DNS and the major DNSBLs itself. The Monday report is keep, rest, or replace. It is not a second health score.