Eight capabilities. Three responsibilities.
Every inbound event is persisted before we acknowledge.
A crash on either side never drops an event. The provider sees a 2xx and stops retrying because we've already written the event to durable storage, signed the receipt, and stamped it with a trace context.
Re-deliver any event, any range, any time.
Pick the events. Pick the destinations. Replay. Pause, resume, cancel — original event identity preserved, idempotency intact. Stop writing one-off scripts at 2 a.m.
Retries on your terms, not the provider's.
Five attempts by default, exponential backoff with jitter. Per-destination policies override the defaults. Inspect, pause, or drain the queue at any time.
One ingest, many destinations.
Route to internal services, third-party APIs, or staging. Per-destination headers, secrets, retry policies. Loopback prevention catches mistakes other tools just deliver.
Know the moment a payload shape changes.
Every payload is fingerprinted and versioned per source and event type. New fields, removed fields, type changes — surfaced before they break consumers. Annotate the fields you care about and they become indexed for search.
Verify what comes in. Sign what goes out.
Built-in profiles for the providers you already use — Stripe, GitHub, Shopify, Adyen, Plaid, Svix — plus custom HMAC for the ones you don't. Every rejection logged with reason.
Trace every event end-to-end.
Filter by source, status, schema version, or any indexed payload field. Inspect headers, body, signature outcome, and every delivery attempt — request and response — for the full lifecycle of any event.
Built on OpenTelemetry.
Traces, structured logs, and audit records into the same pipeline you already run. The platform watching your events can be watched by yours.
Free for the first 10,000 events. No card required.