Triggers
How triggers work: polling schedules, webhooks, and protocol threads.
A trigger is what kicks off an automation. Most automations have one trigger; some have several — a webhook for instant updates plus a polling backstop, or two webhooks for different events from the same app. Claude picks the trigger type (or types) based on what your source system supports and what your prompt describes.
You describe what should kick the run
You tell Claude what event should start the workflow:
> When a new HubSpot contact is created, ...
> Every morning at 8am, ...
> When a webhook from Stripe fires, ...
Claude looks at the source connector, picks the right trigger style, and sets it up. You don't pick "polling" vs "webhook" yourself.
The four trigger types
Four broad types worth knowing:
- Webhook — the source system sends an event to APIANT the moment something happens. Lowest latency. Used whenever the source supports it.
- Polling — APIANT calls the source on a recurring schedule and asks "anything new?" Used when the source doesn't support webhooks.
- DateTime — fires on a clock-based schedule, regardless of what's happening in any external system. Used for nightly reports, periodic reconciliation, and similar timing-driven runs.
- Manual — fires only when a human kicks it off. Used for ops scripts and on-demand work.
You don't have to know which of these Claude picked. If you want to know, ask: "what kind of trigger does this automation use?"
Deduplication
A polling trigger calls the source on a recurring schedule and would see the same record on every tick if APIANT didn't dedup. APIANT's polling triggers automatically deduplicate so a record never fires the automation twice. You don't write any dedup logic.
See also
- How Data Flows — what happens after the trigger fires
- Two-Way Sync — paired triggers across two systems