Copy.ai

Run Copy.ai Workflows from your flows. Agents trigger GTM and copywriting workflows with per-record inputs, wait for or poll run results, and manage completion webhooks for downstream automation.

8 actions API key available

Copy.ai website ↗ Platform Documentation ↗ Capability data verified 2026-07-27

A contact enters the outreach-ready segment in [HubSpot](/integrations/hubspot)

The agent maps the record's fields into Start Variables that match the Copy.ai workflow's configured input names

Run Workflow starts the personalization run and returns its run id

Get Workflow Run polls until the status reads COMPLETE and the generated output is attached

The agent screens the output for unfilled variables and claims that do not appear in the source record

The account owner reads the draft next to the record it was generated from and approves, edits, or rejects it

The approved copy writes back to the contact record and enters the send sequence

What This Integration Enables

Copy.ai is built for teams that have already turned their messaging into Workflows: repeatable GTM plays with named inputs and a defined output, not one-off prompts in a chat box. That structure is exactly what makes it automatable. A FlowRunner agent treats each Copy.ai Workflow as a callable function, feeds it one record's fields at a time, and routes the result to the system where it will actually be used. The judgment calls, whether a draft is safe to send and whether a batch is worth its credits, stay with people through FlowRunner's [human-in-the-loop](/concepts/human-in-the-loop) steps. - Trigger any Copy.ai Workflow with per-record Start Variables pulled from your CRM, sheet, or store - Wait inline for a run to finish, or start it and poll on your own schedule for long generations - Audit run history per workflow to reconcile output volume against credits used - Register completion webhooks so downstream automation reacts the moment a run finishes

Without FlowRunner

Personalization is rationed Custom first lines exist for the ten biggest accounts because nobody has time to write five hundred

Copy moves by hand Output gets pasted from the generation tool into the CRM, the doc, and the email one record at a time

Volume hides errors A broken merge field or a wrong claim ships to a whole segment before anyone reads a single send

With FlowRunner

Every record gets the workflow Agents feed each lead's own fields into the run, so personalization comes from inputs, not from a template

Output lands where it is used Completed runs write back to the CRM, the content workspace, or the sheet without a paste step

A person clears the send Drafts queue for review with their source record attached, and approval is the gate to the sequence

Use Case Scenarios

Outreach personalization at segment scale

A list in [HubSpot](/integrations/hubspot) fills with contacts who hit the qualification threshold. For each one, the agent assembles Start Variables from the contact's industry, role, and recent activity, then calls Run Workflow and Wait against the team's outreach workflow. Each returned draft is attached to its contact and queued in [Slack](/integrations/slack) for the account owner, who approves or edits before the sequence sends. The writing happens at machine speed. The decision about what reaches a prospect's inbox does not.

Product descriptions for the whole catalog

A merchandiser drops three hundred new SKUs into [Google Sheets](/integrations/google-sheets). The agent walks the rows, calls Run Workflow with each product's attributes, and records every run id. A polling loop calls Get Workflow Run until each run reads COMPLETE, then writes the description into the row beside its SKU. The merchandiser reviews the sheet, fixes the handful that miss, and only then does the agent push approved descriptions to [Shopify](/integrations/shopify). List Workflow Runs gives the audit trail when finance asks what the batch consumed.

A content pipeline that reacts to completion

The content team runs long-form generation inside Copy.ai on their own schedule. The agent registers Create Webhook scoped to the blog workflow, so every completed run notifies FlowRunner with the run id, output, and credits used. On each event, the agent creates a draft page in [Notion](/integrations/notion) with the output and its source inputs, then posts the link to the editorial channel in [Slack](/integrations/slack). Editors pick up finished drafts without polling, and nothing publishes until a person moves it out of draft.

Human-in-Loop Highlight

Run Workflow is a spending operation. Every call consumes workspace credits, and the completion payload reports credits used for a reason: a batch launched against five thousand records is spend that cannot be recalled, whether the output is usable or not. So a FlowRunner agent never scales a batch on its own. It runs the workflow against a small pilot slice, posts the sample output, the full segment size, and the credits the pilot consumed, and asks the owner to authorize the rest. The same discipline applies at the other end of the pipeline: generated copy that reaches a customer's inbox cannot be unsent, so the send is a human decision, not a workflow step that happens to follow generation.

Agent processes routinely

Detects exception requiring judgment

Clear match Continues automatically

Ambiguous Routes to human via preferred channel

Human decides

Agent resumes with decision

Agent Capabilities

8 actions

Workflows

Run Workflow Starts a run of a Copy.ai Workflow with a Start Variables object and returns the run id immediately. The run executes asynchronously, which suits large batches where the agent fans out many runs and collects results later. Optional metadata is stored on the run and echoed in completion webhooks.
Run Workflow and Wait Starts a run and polls until it reaches a terminal status, returning the completed output in one step. The right call when a flow needs the copy before it can continue. Poll interval and timeout are configurable, and a timeout returns the last observed state with a timedOut flag rather than failing silently.
Get Workflow Run Retrieves the current status and output of a run started with Run Workflow, including the inputs it was started with and timestamps. The polling step in batch pipelines, and the lookup step when a webhook or a reviewer needs a run's full context.
List Workflow Runs Lists a workflow's runs, most recent first, with id, status, and timestamps. The audit view: reconcile what a batch actually executed, locate a run id after the fact, or spot a failure pattern across a day's generations.

Webhooks

Create Webhook Registers an endpoint that Copy.ai notifies when a workflow run completes, scoped to one workflow or the whole workspace. The completion payload carries the run id, output, echoed metadata, and credits used, which turns Copy.ai into an event source for downstream flows.
List Webhooks Retrieves every webhook registered for the workspace with its target URL, event type, and workflow scope. Used to review what is wired up before adding or removing anything.
Get Webhook Retrieves a single webhook by id with its target URL, event type, and scope. Used to confirm a registration before a flow depends on it.
Delete Webhook Removes a registered webhook so events stop flowing to its endpoint. The catalog is explicit that this cannot be undone, so cleanup flows confirm the id against List Webhooks before calling it.

Frequently Asked Questions

What can FlowRunner do with Copy.ai?

FlowRunner agents can run Run Workflow, Get Workflow Run, and List Workflow Runs in Copy.ai, plus 5 more actions.

Does connecting Copy.ai to FlowRunner require OAuth?

No. Copy.ai connects to FlowRunner with an API key, no OAuth flow required.

Can Copy.ai trigger a FlowRunner workflow automatically?

Copy.ai doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.

Start building with Copy.ai

$100 in credits. No card required. Connect in minutes.

Start Building Free Talk to us