Proof Docs
API referenceOpen Proof

Production monitoring

Continuously replay approved Playwright tests against production without invoking AI on the healthy path.

One dispatcher, durable sweeps

One Vercel Cron request leases every due schedule and starts a durable Workflow execution. Workflow divides the active catalog into bounded chunks and sends each test to the isolated Playwright Function worker. The worker runs Chromium, returns structured results, and never calls a model.

Create a monitor

bash
curl "$PROOF_PUBLIC_URL/api/monitors" \
  --request POST \
  --header "Authorization: Bearer $PROOF_API_KEY" \
  --header "Content-Type: application/json" \
  --data '{
    "repositoryId": "repo_storefront",
    "environmentTargetId": "target_production",
    "name": "daily production",
    "cadenceMinutes": 1440,
    "maxTests": 50,
    "maxParallelism": 4
  }'

The API key determines which organization owns the monitor.

Retry before escalation

Every failure, including an executor timeout, receives one clean deterministic retry. A repeated failure is recorded with durable logs, screenshots, and traces, then sent to the optional signed alert webhook. AI adjudication remains off unless both the global AI switch and the monitor escalation switch are enabled.

Hard spend boundary

Each scenario reserves two cents before dispatch, covering the Function shard and one retry. All Proof billable entry points share a conservative monthly admission ledger capped at $10. Once exhausted, new work is recorded as budget blocked instead of starting compute.