Credential provider
Nango, Auth0, Scalekit
Maintains connections, refreshes tokens and runs provider authorization flows.
RECOVERY HANDOFF
Connection identity and the provider error
Why Revive
Credential reconnect should resume work, not restart it.
Revive coordinates the recovery step between your credential provider and durable workflow runtime.
One recovery path across three systems.
Revive
Recovery control plane
Links the failed connection to the exact run, checkpoint and pending action.
RECOVERY HANDOFF
Verified lease generation and a resume signal
Workflow runtime
LangGraph, Temporal
Persists workflow state, schedules work and resumes the existing execution.
RECOVERY HANDOFF
Run ID, checkpoint and action context
What changes during a real failure.
The difference is visible at the point where a workflow would normally stall, restart or replay an action.
A refresh token is revoked
Without recovery coordination
The worker retries a credential that cannot recover or exits without a reconnect path.
With Revive
Revive classifies the failure, parks the run and sends the account owner a scoped reconnect request.
A different account reconnects
Without recovery coordination
The workflow can resume under the wrong provider identity.
With Revive
The provider subject and tenant must match the original connection before the request is consumed.
The worker times out after a mutation
Without recovery coordination
A blind retry may send the email, create the ticket or submit the payment twice.
With Revive
The action ledger requires reconciliation before an ambiguous mutation can execute again.
The original worker disappears
Without recovery coordination
Recovery context is trapped inside a process that no longer exists.
With Revive
A replacement worker reads the durable case and resumes the same logical run from its checkpoint.
Inspect the full recovery record.
The lab shows the provider signal, saved checkpoint, verified credential generation and action ledger in one timeline.