Summary
A Replit AI agent called an internal function (setupClerkWhitelabelAuth()) without user instruction, which corrupted the Clerk domain sync state for my deployment. Every republish attempt has failed for 6+ hours with the error “Failed to sync Clerk domains.”
Project
Replit app: chipdb.replit.app
Custom domain: nflprospectsdatabase.com
Deployment type: Autoscale
Exact error on every Republish
Failed to sync Clerk domains
What the agent did
During a session, a Replit AI agent called setupClerkWhitelabelAuth() — an internal Replit function, not user-initiated code. This is what introduced the corrupted sync state. The agent should never have had access to call this function on a live production deployment.
What I have already tried
Disconnected nflprospectsdatabase .com custom domain from the deployment
Republished to chipdb.replit.app only — still fails with same error
Re-added nflprospectsdatabase .com — still fails
Checked .replit file — clean, no Clerk entry in [agent] integrations (only stripe:1.0.0)
Checked Clerk dashboard — no obviously incorrect domain entries
Opened a support ticket (#352759) — escalated but unresolved after 6+ hours
What I believe needs to happen
The Clerk whitelabel sync state tied to this deployment needs to be reset or cleared on Replit’s backend. This is not fixable from the project side — the corrupted state lives entirely in Replit’s infrastructure.
Any help on this would be greatly appreciated.
Thanks!