Resources · n8n integration

Autocalls n8n integrationAutomate Autocalls with n8n.

A phone call becomes a step in your workflow. The Autocalls n8n integration exposes 2 operations across 2 resources, plus a trigger that fires when a call ends. The Call resource makes a call, and the Assistant resource lists what an AI assistant can pick up.

Community node This node is not shipped with n8n: it is a community package (n8n-nodes-autocalls) you install from Settings › Community nodes on a self-hosted instance. npm

Try the Autocalls integration on n8nSee the 2 operations

Verified Trustpilot reviews · AI, automation & growth agency

Why automate

What does the Autocalls n8n integration actually do?

The Autocalls node lets a workflow start a phone call handled by an AI assistant, and the Autocalls Call Ended Trigger lets the same workflow pick up where that call stopped. Two resources are covered, Call and Assistant, for a total of 2 operations. Nothing is scheduled inside Autocalls: the call happens because a node ran, and the follow up happens because the call ended.

First scenario, outbound qualification. A new row lands in a spreadsheet, the workflow reads the phone number, and call.make fires the assistant that asks the two or three questions you would ask yourself. The call ends, the trigger wakes a second workflow, and the outcome goes back into Google Sheets next to the lead it belongs to.

Second scenario, internal alerting. Sales teams rarely read a CRM tab in time. Once the call ended payload arrives, a short branch posts the important ones to Slack and leaves the rest alone, so the channel stays readable. Keep the branch condition on data the payload actually carries, not on a field you hope is there.

Third scenario, the written follow up. A call that went well deserves a message while it is still fresh. The trigger hands the result to a formatting step, and Gmail sends the recap. That pattern works because the trigger is a webhook, which means Autocalls calls an n8n URL instead of n8n asking Autocalls every few minutes.

When should you skip the node? The Autocalls API is wider than 2 operations. Anything the node does not expose goes through the HTTP Request node, which calls any endpoint of the tool while reusing the same credential through predefined authentication. That is the normal escape hatch, not a workaround.

Two limits are worth knowing before you build. The node is a community node, so it lives on a self-hosted instance, and n8n Cloud only accepts verified community nodes. And the trigger has no event list: it covers the end of a call, nothing else. If you want a broader picture of the platform before committing, the n8n review covers what it does well and where it hurts.

Connect

How do you connect Autocalls to n8n?

  1. 01

    Install the community package

    Autocalls does not ship with n8n. Open Settings, then Community nodes, and install the npm package n8n-nodes-autocalls, currently in version 1.0.3. This works on a self-hosted instance, whether n8n runs through Docker or npm. On n8n Cloud, only verified community nodes can be installed, so check that point before you plan a Cloud rollout. Once installed, Autocalls appears in the node panel like any built in node.

  2. 02

    Create the Autocalls credential

    A credential is created once from the Credentials menu and reused in every workflow that needs it. The node has no authentication selector: it simply offers the credential in a dropdown, and you pick it. Create it before opening your first workflow, it saves a round trip. The same credential also serves the HTTP Request node through predefined authentication, for the Autocalls endpoints the node does not cover.

  3. 03

    Pick an assistant and run the node once

    Drop the Autocalls node on the canvas and run the Get assistants operation on its own. It returns the assistants of your account, which gives you their identifiers without leaving n8n. Then switch to Make a phone call: the Assistant Name or ID field offers the list, or takes an identifier through an expression such as {{ $json.assistantId }} when the choice comes from an earlier node. One run on your own number confirms the credential works.

Triggers

What starts an Autocalls workflow?

Autocalls Call Ended Trigger is the node that starts a workflow each time Autocalls receives a new submission. Pick the form, activate the workflow: n8n registers the webhook on your Autocalls account.

What you see in n8n

Configuration notes

01Autocalls Call Ended TriggerThis trigger runs a workflow when an Autocalls phone call ends. It is a webhook, which means Autocalls calls an n8n URL as soon as the call is over and the execution starts within seconds. There is no polling interval to tune and no event list to tick: the end of a call is the only thing it listens to. The node is in version 1.

This trigger runs a workflow when an Autocalls phone call ends. It is a webhook, which means Autocalls calls an n8n URL as soon as the call is over and the execution starts within seconds. There is no polling interval to tune and no event list to tick: the end of a call is the only thing it listens to. The node is in version 1.

Key parameters

  • Assistant Name or ID: required, and the only field to fill. The dropdown lists the assistants of your account, or you pass an identifier through an expression when the assistant is chosen upstream.
Use cases
one workflow per assistant is usually cleaner than one workflow filtering everything, because each assistant has its own script and its own follow up. A qualification assistant feeds a scoring branch, a reminder assistant feeds a simple log.
02Register the webhook URL in AutocallsMost triggers in n8n register themselves with the tool when the workflow is activated. This one does not. The URL that n8n generates has to be copied into the assistant settings inside the Autocalls dashboard and saved there, by hand, once per assistant. Nothing in n8n warns you if you forget, the workflow just never runs.

Most triggers in n8n register themselves with the tool when the workflow is activated. This one does not. The URL that n8n generates has to be copied into the assistant settings inside the Autocalls dashboard and saved there, by hand, once per assistant. Nothing in n8n warns you if you forget, the workflow just never runs.

When to use it
do this pass right after building the trigger node, while the URL is still on screen. Test URLs and production URLs differ in n8n, so the one you paste in Autocalls is the production one, and the workflow has to be active for it to answer. If you duplicate a workflow later, the new copy gets a new URL and needs the same manual step again.
Actions

What can the Autocalls node do?

The Autocalls node exposes 2 operations across 2 resources. For each one: the node as you configure it in n8n, the required fields, and our field notes.

Resources × operations matrix
ResourceGet AssistantsMake
Call✓
Assistant✓

Call

1 operation
01

Make a phone call

call.make

What you see in n8n

Notes & use cases

Starts a phone call to a customer, handled by one of the AI assistants of your Autocalls account. The workflow does not wait for the end of the call, which is why the trigger exists.

Key parameters

  • Assistant Name or ID: required, the assistant that will run the conversation. Pick it in the list, or pass an identifier with an expression when several assistants share one workflow.
  • Customer Phone Number: required, the number to reach, most often {{ $json.phoneNumber }} coming from the node before.
  • Variables: optional, the values passed to the assistant so the script can name the customer or the product instead of staying generic.
Use cases
calling back a form submission within the minute.

Assistant

1 operation
02

Get assistants

assistant.getAssistants

What you see in n8n

Notes & use cases

Returns all the assistants of your Autocalls account. No field to fill, which makes it the fastest way to check that the credential is valid right after installing the node.

Use cases
two of them stand out. Building an internal list, so a colleague picks an assistant from a form instead of pasting an identifier. And auditing: run it before a campaign to confirm the assistant you are about to call still exists under the name you wrote in the workflow. It also answers the question the other operation raises, since Make a phone call expects an assistant identifier and this one hands it over.
Need help

Need help automating Autocalls with n8n?

A person reads every message.

FAQ

Autocalls and n8n, the questions that come next

01Is the Autocalls n8n integration free?
Yes on the n8n side. Autocalls is not shipped with n8n: it is a community node, published as the npm package n8n-nodes-autocalls in version 1.0.3, and installing it adds no cost to your n8n bill. You add it from Settings, then Community nodes, on a self-hosted instance, and n8n runs the same whether you installed it through Docker or npm. One restriction matters: n8n Cloud only accepts verified community nodes, so a Cloud instance may simply refuse it. What the Autocalls platform itself charges for calls is a separate topic and this page does not cover it.
02What credentials do you need to connect Autocalls to n8n?
One Autocalls credential, created once. In n8n a credential is the stored authentication for a tool: you create it from the Credentials menu, and every workflow that uses the Autocalls node then picks it from a dropdown instead of holding secrets of its own. The node has no authentication selector, so there is no method to choose between: one credential type, one dropdown. The same credential is also what the HTTP Request node reuses through predefined authentication when you need an Autocalls endpoint the node does not expose. The exact fields to fill live in the Autocalls documentation.
03What are the limits of the Autocalls node in n8n?
The node covers 2 operations across 2 resources: Make a phone call, on the Call resource, and Get assistants, on the Assistant resource. That is a deliberate scope, not a bug. Everything else the Autocalls API offers goes through the HTTP Request node, which calls any endpoint of the tool and reuses the same credential through predefined authentication. The trigger has the same narrow shape: it reacts to the end of a call and offers no list of events to choose from. And because it is a community node, it needs a self-hosted instance unless the node has been verified for n8n Cloud.
04Does the Autocalls trigger react in real time?
Yes, it is a webhook trigger, so the execution starts within seconds of the call ending. A webhook here means Autocalls calls a URL that n8n generated, instead of n8n asking Autocalls every few minutes whether something happened. There is no interval to configure and no delay to absorb. One manual step stands between you and that behavior: the webhook URL is not registered automatically. Open the assistant settings in your Autocalls dashboard, paste the URL n8n gave you, and save. The workflow also has to be active, because a production webhook URL only answers on an active workflow.
05n8n or Make for Autocalls?
It depends on where you want the workflow to live. Make is a hosted automation platform with no self-hosting option, billed per operation, and its visual logic suits linear scenarios well. n8n runs on your own server or on the hosted offer from n8n, and a workflow is identical in both cases, which matters here: Autocalls is a community node, and installing community nodes is a self-hosted move unless the node is verified for Cloud. Data control follows the same line, since a self-hosted instance keeps call payloads on your infrastructure. On cost, a per operation model and a server you already pay for do not compare the same way once volume grows.
Same platform

Other docs you might need on this stack.

Browse all integrations