Mittr — Webhook infrastructure you don't have to build.

Built to send.<br>Send the event.We do the rest.

You'd build the queue. The retries. The audit log. The dashboard that explains the silent failures. We already did.

Start free, no credit card Read the docs<br>Every event persisted Automatic retries Full audit trail

POST /api/v1/events

order.created<br>12:00:00<br>queued · POST /api/v1/events

Attempt 1<br>+1.2s<br>502 Bad Gateway<br>retry in 1.2s

Attempt 2<br>+3.2s<br>504 Gateway Timeout<br>retry in 4.8s

Attempt 3<br>+9.8s<br>200 OK · delivered

3 attempts · 9.8 s · auto-recovered without your code lifting a finger

For developers One call. Same delivery guarantees in every language.<br>The same wire format the SDK, the docs, and the dashboard all agree on. Idempotent retries, signed delivery, and a full audit trail — without writing any of that yourself.

Node Python Go curl<br>import { Mittr } from "@mittr/sdk-node";

const mittr = new Mittr({ apiKey: process.env.MITTR_API_KEY });

await mittr.events.send({<br>eventType: "order.created",<br>payload: { orderId: "ord_123", amount: 9900 },<br>});<br>import requests

requests.post(<br>"https://app.mittr.io/api/v1/events",<br>headers={"X-API-Key": "mtr_your_key"},<br>json={<br>"eventType": "order.created",<br>"payload": {"orderId": "ord_123", "amount": 9900},<br>},<br>body, _ := json.Marshal(map[string]any{<br>"eventType": "order.created",<br>"payload": map[string]any{"orderId": "ord_123", "amount": 9900},<br>})<br>req, _ := http.NewRequest("POST", "https://app.mittr.io/api/v1/events", bytes.NewReader(body))<br>req.Header.Set("X-API-Key", "mtr_your_key")<br>http.DefaultClient.Do(req)<br>curl -X POST https://app.mittr.io/api/v1/events \<br>-H "X-API-Key: mtr_your_key" \<br>-H "Content-Type: application/json" \<br>-d '{<br>"eventType": "order.created",<br>"payload": { "orderId": "ord_123", "amount": 9900 }<br>}'<br>Copy

@mittr/sdk-node is on npm — typed client, signature verification, framework-agnostic webhook middleware. Read the full quickstart →

For AI agents Give your agents a delivery layer that doesn't forget<br>Mittr speaks MCP. An agent calls it as tools — so its side effects get the same retries and audit trail as the rest of your traffic, not a fire-and-hope HTTP call.

Your agent drives Mittr as tools<br>Connect over MCP and an agent can set up endpoints, send events, and inspect or replay deliveries — the same operations your code gets, exposed as tools it can call.

Agent actions become durable<br>A side effect stops being a fire-and-hope HTTP call. Every send gets the same backoff, circuit breakers, and dead-letter recovery as the rest of your traffic.

Every run is auditable and replayable<br>Events are correlated by agent run and fully logged — so you can see exactly what an agent did, and replay any delivery when it needs a second attempt.

Read the AI agents guide →

The problem Webhooks fail in ways you won't see coming<br>You've shipped webhook delivery before. It worked fine until it didn't.

Events vanish silently<br>Server blips eat events. No retry. No log. No trace. You find out when the customer emails.

Retries are an afterthought<br>You bolt on a retry loop. It hammers failing endpoints with no backoff, no circuit breaker. Dead events pile up until recovery becomes a weekend project.

Zero visibility into failures<br>Something breaks. You hear about it from customers, not your systems. No delivery history. No status codes. No latency data. Nothing to debug with.

The solution Mittr delivers. Then proves it.<br>Every event persisted before delivery. Every attempt logged. Every failure recoverable — by you, from the dashboard.

Every event is persisted first<br>Your event is written to durable storage before we attempt delivery. If our infrastructure crashes mid-flight, your event survives. Transactional outbox pattern, visible in your dashboard.

Smart retries with circuit breakers<br>Exponential backoff per endpoint. Circuit breakers stop hammering failing destinations. Dead events can be edited and resent — not just replayed blindly.

Full delivery observability<br>Every attempt logged with status code, latency, and response body. Alert rules notify you on Slack or PagerDuty before your customers notice anything broke.

How it works From signup to first delivered event: five minutes<br>Three API calls. One dependency. Production-grade infrastructure.

01<br>Register your endpoint<br>Tell Mittr where to deliver. Set event types, retry policy, and timeout. One API call.

02<br>Send events<br>POST your payload with an idempotency key. Mittr handles fan-out, signing, and delivery.

03<br>Ship, don't babysit<br>Monitor from the dashboard. Replay failures. Edit and resend dead events. Every delivery is accounted for.

Capabilities Everything you need. Nothing you don't.<br>Every feature available on every plan. Only numeric limits differ.

Fan-out routing<br>One event delivered to every matching endpoint. Subscribe by type with wildcards.

Security: signing, mTLS, allowlists<br>HMAC-SHA256 on every delivery. Zero-downtime secret rotation. Mutual TLS with client certificates. CIDR allow/block lists per...