Push KOLens growth alerts to Slack, your inbox, or any webhook URL

The alerts feed nobody visits

Growth signals only matter if you act on them while they're fresh. A "rising KOL" alert that you discover three days later isn't a signal — it's archaeology. KOLens has shipped a detector + an in-app feed at /alerts for a while, but the loop was always: notification email → log into KOLens → check feed. That first step is the one most users skipped.

The new delivery layer cuts the loop. The detector fires → KOLens pushes the alert directly to your inbox or webhook → you click the link in the message and land on the relevant KOL.

How to wire it up

1. 1

   Open /alerts/settings.

   New page. Two cards (Email, Webhook), each with an enable toggle + the contact field, plus mute-by-type checkboxes and a per-hour throttle.
2. 2

   Email — paste an address, save.

   KOLens uses plain SMTP — set SMTP_HOST / SMTP_USER / SMTP_PASSWORD / SMTP_FROM on the API service (Resend, Postmark, AWS SES, Mailgun all expose port 587 + STARTTLS). When SMTP isn't configured the email channel silently no-ops; the settings UI still lets you save the address.
3. 3

   Webhook — paste any HTTPS URL.

   KOLens POSTs the JSON payload (see below) with a 3 second timeout. Slack incoming hooks, Zapier catch hooks, n8n, Tines, your own internal endpoint — anything that takes a plain JSON POST works. Optional shared-secret token sent as X-Kolens-Signature for receiver-side verification.
4. 4

   Send a test, verify, save.

   The "Send test alert" button fires a synthetic alert through every enabled channel and reports per-channel success right on the page. Last-error banner per channel surfaces SMTP / DNS / 4xx failures so you can fix them without grep-ing logs.

Webhook payload shape

Stable JSON. We add fields, never rename — Zaps + scripts that depend on these keys today will keep working as new alert types land.

• id — the alert UUID. Doubles as a Slack message dedupe key.
• type — one of rising_kol, trending_video, audience_drift.
• subject_username / subject_video_id — what the alert is about.
• score — numeric severity (% growth for rising_kol, view spike multiplier for trending_video, percentage-point delta for audience_drift).
• payload — type-specific facts. For rising_kol: { delta_pct, baseline, current }; for audience_drift: { country, delta_pct, prev_pct, new_pct }.
• link — pre-built deep link into KOLens (e.g. https://kolens.ai/kols/minicaseco) so the recipient can jump directly to context.
• created_at — ISO 8601 timestamp.

The per-hour throttle

A trending video can produce a burst (multiple watched KOLs all commenting on it within an hour). Without a throttle that means 15 Slack pings in one minute. The default cap is 20 deliveries per hour per user, shared across all enabled channels — beyond that, alerts still write to the in-app feed but skip push. Adjustable on the settings page (set 0 to mute push entirely without losing the in-app trail).

vs the alternatives

• Modash + Zapier — the indirect path. Modash does the detection, Zapier polls + fires. Two subscriptions + one polling delay. KOLens delivers at detection time natively.
• Roll-your-own Apify scheduler — possible if your team has the engineering bandwidth. KOLens skips the "did it already fire today?" dedup logic that's annoying to get right at scale.
• The old in-app alerts page — still there, now with a "Delivery settings" link in the corner. The push channels are additive, not a replacement.

Next up

AS-3-style watchlists for Facebook Ad Spy will start emitting alerts on the same channel ("CompetitorX just published 3 new ads"). Watch the changelog or wire a webhook today and you'll get the new event types automatically.