Webhooks Twitter Integration

AyeWatch is an AI internet monitor for public pages and topics. For webhooks Twitter integration, it turns an X search or profile into signed JSON events that your service can verify and route.

Three integration shapes

A chat relay posts the match into Slack, Discord, or Teams. A queue puts the event in front of a worker that can retry. A database writes a row for reporting and later review. Each shape has a different failure mode.

Chat is the fastest to build and the easiest to lose. A queue survives a downstream outage. A database is the right choice when the record matters more than the immediate ping.

Verify before you parse

Read the raw body, compute an HMAC-SHA256 with your webhook secret, and compare it to the X-AyeWatch-Signature header in constant time. Reject anything that does not match before JSON parsing.

Store the secret outside the repository, rotate it on a schedule, and keep an overlap window so old and new secrets both verify during the change.

Map the payload once

Keep the raw JSON and a normalized record: update_id, topic_id, author, text, URL, and received time. A schema version field saves you when the payload grows.

Do not fetch the post on every event unless you need media. The text and URL are enough for routing, search, and most notifications. Fetch lazily when a human opens the record.

Route by topic_id

Use one endpoint for all topics and switch on topic_id. A brand mention goes to support, a competitor launch goes to product marketing, and a hiring post goes to recruiting. The filter belongs upstream, in the topic, not in a pile of conditionals.

One endpoint is easier to secure, monitor, and rotate than five endpoints with five secrets.

Design for failure

Acknowledge with a 2xx in a few seconds, then enqueue the work. Retries can arrive out of order, so use update_id as an idempotency key and a received timestamp for ordering.

Add a dead-letter queue and an alert when deliveries fail repeatedly. A silent webhook is worse than no webhook because the team assumes it is working. The Twitter webhook alerts guide covers the setup.

Wire X posts into your system

1. Choose the target shape: chat, queue, or database.
2. Create the X search or profile topic and add the filter.
3. Build an HTTPS endpoint that verifies the HMAC signature.
4. Normalize update_id, topic_id, author, text, URL, and time.
5. Queue slow work, add a dead-letter path, and alert on repeated failures.

The bottom line

Pick one integration shape and build the smallest version: verify, normalize, route. Send a real X match through it before you add a second topic. The pipeline that survives a duplicate and a failed delivery is the one worth expanding.