Zapier
ProductivityBridge FlowRunner into Zapier through the Webhooks by Zapier Catch Hook trigger. A flow posts data to a Catch Hook URL, so agents reach the long tail of consumer apps Zapier already lists.
What This Integration Enables
FlowRunner runs the orchestration. This connector is a deliberately thin outbound bridge for the cases where the last hop is easier to reach through a Zap than to build directly: a niche app somebody on the team already wired up, a legacy system that emits XML, a downstream action a non-technical colleague wants to be able to change without opening a flow. A flow posts to a Webhooks by Zapier Catch Hook URL, the Zap behind that URL runs, and the long tail of apps Zapier lists becomes reachable without moving the orchestration layer out of FlowRunner.
Being clear about the trade is more useful than pretending there is not one. Zapier lists a larger app directory than FlowRunner does, and for a genuinely obscure destination the bridge is the faster path. The mainstream few hundred services most teams actually run on are ones FlowRunner already covers directly and in depth, and a direct connector gives you things a bridge cannot: the run is one execution rather than a chain of separately metered task steps, failures surface in FlowRunner's own run history, and the agent can pause mid-run for human-in-the-loop approval and resume with full context. Everything past the Catch Hook URL is invisible to the flow. Use the bridge for the tail and build the trunk natively. For the full side-by-side, see FlowRunner vs Zapier.
Three details shape how you design with it. This is one-way and outbound, so no account connection exists: the Catch Hook URL is itself the credential and should be treated as a secret. A success response means Zapier accepted the request, not that the Zap finished, and a Zap that has been turned off returns a 404. And Zapier parses JSON, XML and form-encoded bodies only, flattening nested JSON with a double underscore so that a nested user object arrives as user__name, unless Parse Nested Data is turned off.
Without FlowRunner
With FlowRunner
Use Case Scenarios
Reaching a downstream app nobody wants to build for
A finance flow needs to drop a line into a small industry-specific tool that one team uses and that publishes an API nobody wants to own. Someone has already built a Zap for it. The agent finishes its reconciliation work in QuickBooks Online, assembles the payload, and calls Send Data to Zap. The judgment, the matching and the exception handling all happened in FlowRunner, where they can be audited. The last two inches of plumbing stay in the Zap, where the person who cares about that tool can maintain them.
Handing a whole batch across in one request
An agent finishes qualifying overnight leads and has 60 records to pass to a Zap that writes them into a niche CRM. Rather than looping and making 60 outbound calls, it calls Send Multiple Items to Zap once with the array, and Zapier triggers the Zap once per object. The flow logs the item count it sent, so a discrepancy in the destination the next morning is a question with a starting number rather than a mystery. Very large batches get split across several calls to stay inside the request size limit.
Feeding a legacy payload through untouched
An older system emits form-encoded callbacks that a Zap already knows how to interpret. The agent receives the callback, validates it, decides whether it should be forwarded at all, and then calls Send Raw Body to Zap against a Catch Raw Hook trigger so the body and headers arrive exactly as they were sent. FlowRunner does the deciding, which is the part that needs an audit trail and an escalation path. Zapier does the parsing it was already doing.
Human-in-Loop Highlight
Send Multiple Items to Zap is the operation with a real cost attached, and the cost is not obvious from the step. Zapier triggers the Zap once for every object in the array, so a single call with 4,000 records starts 4,000 runs, each one consuming the Zap's own metered task budget and each one potentially sending something outward to a customer. Once the request is accepted, none of it can be recalled from FlowRunner, and FlowRunner cannot even observe the runs to know how far they got. A success response only says Zapier took the request.
The one place that is still controllable is before the post, and FlowRunner is the side that knows the array length. So the agent checks the count against the envelope it expects for that run, and when it is outside, it stops. It messages the operations owner in Slack: "The nightly lead handoff has 4,212 items. The last thirty runs averaged under 200. Each item starts a separate Zap run and sends an outbound email. Release the batch, release the first 200 only, or hold it for review?" The operator decides. This is the digital andon cord applied to the exact point where a flow stops being able to see or undo its own effects: the agent notices that this run does not look like the others, and asks before it crosses the line rather than after.
Agent Capabilities
3 actionsZaps
3- Send Data to Zap Sends a JSON object to a Zapier Catch Hook URL, starting one run of the Zap behind it. Used as the final outbound step when the destination is easier to reach through an existing Zap than to connect directly. Zapier parses the body into individual fields the Zap can map, flattening nested objects with a double underscore unless Parse Nested Data is turned off. An entirely empty request is ignored, so always include at least one field, and remember that a success response means the request was accepted rather than that the Zap completed.
- Send Multiple Items to Zap Sends an array of JSON objects to a Catch Hook URL, and Zapier triggers the Zap once for every object in the array. Used to hand a batch of records across in one request instead of looping, with each object parsed into fields exactly as a single payload would be. The combined request counts against the same size limit as a single payload, so split very large batches across several calls, and treat the item count as the number of downstream runs you are starting.
- Send Raw Body to Zap Sends an already-serialized request body to a Zapier hook URL without FlowRunner reshaping it. Used with a Catch Raw Hook trigger when the Zap needs the untouched body and request headers, or when the data is XML or form-encoded rather than JSON. Zapier accepts JSON, XML and form-encoded content only, and the raw hook carries a smaller size limit than a parsed Catch Hook.
Frequently Asked Questions
What can FlowRunner do with Zapier?
FlowRunner agents can run Send Data to Zap, Send Multiple Items to Zap, and Send Raw Body to Zap in Zapier.
Does connecting Zapier to FlowRunner require OAuth?
No. Zapier connects to FlowRunner with an API key, no OAuth flow required.
Can Zapier trigger a FlowRunner workflow automatically?
Zapier doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.
Start building with Zapier
$100 in credits. No card required. Connect in minutes.