Zoom Technical Support closed my ticket as “configuration-related” and
directed me to this forum, so I’m hoping someone can spot the configuration
fault — I’ve exhausted every remediation previously prescribed on this forum
for the symptom.
**Setup**
- Account-level OAuth app (unpublished), client ID `CRekx91zSHulDOz3UCO0A`
- Event Subscription with `recording.transcript_completed` and
`recording.completed` ticked
- Event notification receiver: **All users in the account**
- Notification endpoint: an HTTPS URL on Vercel that **passes Zoom’s own
URL validation** in the app console (correct `encryptedToken` HMAC)
- OAuth works: token exchange and refresh succeed, and REST calls
(`/users/me`, recordings) succeed through this same app right now
**Symptom**
Five cloud recordings made over two days, all fully processed (Zoom emailed
completion confirmations, transcripts exist). **Zero webhook deliveries of
any kind, ever** — including `recording.completed`, which fires on
processing and doesn’t wait for transcription.
I can be precise about “zero”: the receiving endpoint logs every request it
gets, *including* ones it rejects (bad signature, stale timestamp, missing
signature). The log is empty. Nothing is arriving at the endpoint, which is
the same URL the console validates successfully.
**Remediations already performed, in order**
1. Event subscription **deleted and recreated from scratch** (not edited),
both events ticked, endpoint re-validated, saved
2. OAuth install **removed and re-created after** the rebuilt subscription
existed
3. Fresh cloud recording → processed and confirmed by email → no delivery
4. Install re-created again **via the app’s own “Add App Now” link** (fresh
token confirmed saved 19:16 UTC, 27 Aug 2026)
5. Another fresh cloud recording immediately after → no `recording.completed`
delivery
**Questions**
1. Can someone with access check the **server-side event delivery log** for
app `CRekx91zSHulDOz3UCO0A`? Whether Zoom attempted delivery at all (and
to what URL, with what response) settles this in one look.
2. What configuration state produces *validation succeeding* but *zero
delivery attempts*? A wrong secret or URL would show up as failed
deliveries — this is the absence of attempts.
3. If the app has separate Development and Production environments: which
environment’s install actually receives event deliveries for an
unpublished app, and can a mismatch between the installed environment and
the subscription’s environment produce exactly this?
Happy to share account ID, meeting UUIDs, and exact recording timestamps by
DM.