Social media monitoring. A webhook when they post.
Watch supported accounts, public channels and searches on six platforms. We check on your schedule and send a signed webhook when we find new posts.
What you get
No cron. No scraper. Just events.
Each monitor is a source, a schedule and a webhook. We run the checks, work out what's new, and deliver it.
Only what's new.
Each check runs the lookup, compares it with the last one, and sends an event only when something new appears.
- A baseline on creation, so day-one content never fires
- Run on an interval, or on a cron with a timezone
- Events use the same item schema as the endpoint
Test without deploying.
See a real, signed delivery before you write a handler. No public URL, no tunnel.
- A hosted test inbox that captures deliveries
- socialfetch listen forwards events to localhost, byte for byte
- Trigger a check or send a sample event on demand
Deliveries you can count on.
If your endpoint is down, we retry 8 times over about 21 hours, and log every attempt's headers, payload and response.
- One-click redelivery from the log
- Endpoints switch off after 3 days of failures, or at once on HTTP 410
- Monitors keep recording events while an endpoint is down
Signed like Stripe
HMAC-SHA256 over the raw body in Stripe's t=…,v1=… format, with secret rotation that never drops a delivery.
Cost before you commit
A dryRun create returns the credits per check, per day and per month for your schedule, before anything is created.
30 days of events
Pull any event from the API, or redeliver any attempt, for 30 days after it happened.
How it works
Your first webhook in minutes.
Three calls, or a few clicks in the dashboard.
Step 1
Pick a source and a schedule
GET /v1/monitors/sources lists everything you can watch, its minimum interval and a sample event.
Step 2
Preview the cost, then create
dryRun=true returns the credit math. Create for real and we record a baseline, so old posts don't fire.
Step 3
Verify and go live
Check the signature against the raw body with the SDK, then point the monitor at your real endpoint.
Pay per check. Nothing else.
A check costs the same credits as the lookup it runs, whether or not it finds anything new. No monitor fee.
Plan limits cap how many monitors you run, not how often they check. A dryRun shows the exact cost for your schedule before you create anything.
Cost preview
in credits- Setup creditsCreating, editing, testing and triggering are free
- 0
- Credits a checkFor example, an X account's latest posts
- 2
- Free creditsEnough to see your first webhooks land
- 100
No separate subscription. Pause or stop at any time.
Everything is an API call
Monitors and webhook endpoints are part of the same API as every lookup, with the same key, SDK and docs.
Monitors docs
- Monitoring quickstart
Create a monitor and receive your first webhook end to end.
- Testing webhooks locally
The hosted test inbox, socialfetch listen and manual triggers.
- Delivery, retries & idempotency
The retry ladder, at-least-once delivery and deduping.
Common questions
What counts as a billed check?
Every scheduled or manually triggered poll, billed at the underlying operation's normal per-request credit price — whether or not it finds anything new. There's no discount for an empty result and no separate monitor fee.
Do monitor or webhook management calls cost credits?
No. Creating, listing, getting, updating, deleting, pausing, resuming, triggering, and testing are always free. Only the scheduled or triggered poll itself bills.
What happens if my endpoint goes down?
Failed deliveries retry automatically across 8 attempts spanning roughly 21 hours. The endpoint auto-disables after 100 consecutive failures or 3 days of continuous failure (or immediately on HTTP 410), but your monitor keeps polling and recording events the whole time — pull them via the events API or redeliver manually once you're back up.
Can I test this without deploying a webhook endpoint?
Yes. A hosted test inbox (sink endpoint) captures real, signed deliveries with no outbound URL and no deploy, and socialfetch listen forwards those same signed events to a server running on your machine.
Which platforms can I monitor today?
Twitter/X, Instagram, TikTok, Reddit, Bluesky, Threads, Telegram and YouTube. The sources endpoint lists the supported account, channel and search options. New sources use the same schedules, events and webhook delivery.
Is delivery exactly-once?
No — at-least-once, like every webhook system worth using. Retries and manual redeliveries carry the same event id as the original, so dedupe on that (not on the per-attempt delivery id) and your handler stays correct regardless of how many times an event actually arrives.
Watch your first account.
100 free credits. Create a monitor, see a real signed webhook land, and verify it end to end.
No credit cardSign up in 30 seconds withGoogleGitHub