LinkedIn outreach with n8n.
n8n LinkedIn outreach with our tool runs on plain HTTP in both directions: our webhooks send events such as message.replied to an n8n Webhook node, and an n8n HTTP Request node adds leads to a campaign through our API. We have no native n8n node, and n8n’s own LinkedIn node only creates posts.
Table of contents
Before you start
- An n8n instance that can receive requests from the internet, cloud or self-hosted, so our webhooks can reach its production URL
- A workspace with at least one launched campaign and one connected sender
- An API key from Settings → API, kept in n8n as a credential and never pasted into a workflow node
The steps
Know what each side does
n8n’s documentation (read 2026-11-15) lists one operation for its built-in LinkedIn node: create a post, as a person or an organization. It sends no connection requests and no messages, so it is not an outreach tool on its own.
We don’t publish an n8n node. What we have is an API and webhooks, and n8n can call and receive both through its Webhook and HTTP Request nodes. Everything below is built from those two nodes.
The split is simple. Our app sends the connection requests and messages from each sender, inside that sender’s limits and working hours. n8n moves information around them: replies out to your CRM or a chat channel, new leads in from wherever you find them.
n8n never sends from a LinkedIn account in this setup. The senders you connected do, at their own limits, whatever n8n adds.
Create an API key and store it in n8n
In our app, open Settings → API and create a key. It is one key per workspace, and you can revoke it later.
In n8n, create a credential of the Header Auth type, with the name X-API-Key and your key as the value. n8n’s HTTP Request node documentation (read 2026-11-15) lists Header auth among its generic authentication types, so the credential is picked in the node and stays out of the workflow itself.
Our API accepts the key in the X-API-Key header. Every call below goes to the /public/v1 path on our API host.
The API is rate limited per key. If a workflow fires many calls at once, use the HTTP Request node’s batching options, which n8n documents as items per batch and a batch interval, to space them out.
Receive replies: add a Webhook node in n8n
Create a workflow that starts with n8n’s Webhook node. n8n’s documentation (read 2026-11-15) says the node accepts POST among other methods, shows a test URL while you listen for a test event and a separate production URL once the workflow is published, and can require Header auth.
Set the HTTP method to POST, choose Header auth with a new credential, and give it a name and a long random value of your own, for example the name X-Webhook-Token. Set the response mode to Immediately, so n8n answers 200 as soon as it has the event.
Copy the production URL when you are ready to go live. Use the test URL first.
Our webhooks retry a failed delivery with backoff, up to 8 times over 24 hours. A workflow that is unpublished or answers with an error will make us retry the same event.
Point our webhook at n8n
In our app, open Settings → Webhooks and add an endpoint with the n8n URL. Add one custom header with the same name and value you gave n8n’s Header auth, so n8n can reject anything else. We allow up to 5 custom headers per endpoint, and credentials go in a header, never in the URL.
Pick the events you want. For a reply alert that is message.replied, which fires on a lead’s first reply, and optionally message.received for every later message. The full list is connection.sent, connection.accepted, message.sent, message.replied, message.received, lead.tagged, lead.finished and campaign.finished.
Optionally filter the endpoint to one campaign, so an n8n workflow for one team’s campaign never sees another’s.
Settings lists every delivery we made to the endpoint, so a missing event can be checked on our side first.
Read the payload and act on it
Each delivery is JSON. It carries the event name, a timestamp and a data object. For message.replied, data holds the lead (name, company, title, LinkedIn URL), the campaign, the sender and the message body.
Send a real reply from a test account, or one from a colleague, and open the test execution in n8n. The node’s output shows the exact shape, so build your next nodes from what you see rather than from a guess. From there, any n8n node can take the data: post to a chat channel with the lead’s name and the reply, create a CRM record, or add a row to a sheet.
Every delivery also carries X-Event, X-Delivery-Id, X-Timestamp and X-Signature headers. The signature is sha256= followed by the hex HMAC-SHA256 of the timestamp, a full stop and the raw body, using the endpoint’s signing secret. If you want to verify it, do it in a Code node. The custom header already keeps strangers out, so this is a second check, not a requirement.
A reply stops that lead in every campaign and lands in the unified inbox. The webhook is a notification. Answering still happens in the inbox, or through the conversations API.
Add leads from n8n
To feed a campaign from n8n, add an HTTP Request node after whatever finds the leads, with method POST and the Header auth credential from step 2. First list your campaigns with GET /public/v1/campaigns, which returns each campaign’s id, name and status, and note the id of the one to feed.
Then post to /public/v1/campaigns/{campaign_id}/leads with a JSON body:
{
"leads": [
{
"profile_url": "https://www.linkedin.com/in/example",
"first_name": "Priya",
"last_name": "Rao",
"company": "Tallgrass",
"title": "VP Sales",
"custom_fields": { "event": "Austin meetup" }
}
]
}
You can send 1 to 100 leads per call. The response says how many were added, how many were skipped and how many had an invalid profile URL. A lead that can’t be enrolled, such as one already live in another campaign, is counted as skipped. The campaign has to be launched: a campaign still in draft answers with an error.
Leads sent this way also join the campaign’s list, and each one goes to the connected sender with the most weekly room left.
Set a fallback on every variable the messages use, such as {first_name|there}, because an API lead can arrive with a field empty.
Stop, tag and pause from n8n
The same API key does the other jobs n8n is good at.
POST /public/v1/campaigns/{campaign_id}/stop-leadwith{"profile_url": "..."}takes one lead out of a campaign, for example when your CRM shows they are already a customer.PUT /public/v1/leads/tagswith a profile URL and a list of tag names replaces that lead’s tags. Missing tags are created.GET /public/v1/leads?profile_url=...reads a lead with its lists, campaigns, state and tags.POST /public/v1/campaigns/{campaign_id}/pauseand/resumepause or resume a campaign.GET /public/v1/conversations?since=...returns threads that changed since a time, andPOST /public/v1/conversations/{id}/replyreplies as the conversation’s sender.
A reply sent through the API goes out as the conversation’s sender, so treat it like a person writing: a template reply to every answer reads as one.
Plan for a team, not one account
n8n can add leads faster than any account can contact them. It makes no difference to the pace. Each connected sender starts at 15 connection requests a day, can be set up to 25, and never sends more than 100 a week, and it works Monday to Friday, 09:00 to 18:00 in its own time zone. Messages start at 40 a day, up to 80.
Across 5 senders that is 75 requests a day to start, up to 125, and 500 a week at most. A list of 2,000 leads takes weeks either way. See how the weekly invitation limit works and how many connection requests a team can send.
The price is $29 a month per connected sender, with every feature and unlimited free members: $29 for 1 sender, $87 for 3, $145 for 5 and $290 for 10. The API and webhooks are on every plan. There is no per-workflow or per-call charge from us, and n8n has its own pricing. See pricing.
One workspace, one API key, one webhook endpoint per team workflow is enough for most teams. Use a campaign filter to keep them apart.
A reply to any sender’s message arrives in an n8n workflow as a webhook event, and an n8n workflow can add, tag and stop leads in a campaign through our API.
Is there an n8n LinkedIn automation node for outreach?
Not from us. n8n’s built-in LinkedIn node, per its documentation read 2026-11-15, has one operation, which creates a post. For outreach, n8n’s Webhook and HTTP Request nodes talk to our webhooks and API directly. We don’t publish an n8n node or a community node.
How do I run LinkedIn automation with n8n?
Keep the sending in our app, where each connected sender has its own limits and working hours, and use n8n for the surrounding work. A Webhook node receives events such as message.replied. An HTTP Request node adds leads to a campaign, tags them or stops them through the API.
Can n8n send LinkedIn connection requests or messages by itself?
Not through n8n’s own LinkedIn node, which only creates posts. With our tool, the senders you connected send the requests and messages. n8n can add leads to a campaign, and it can reply to a conversation through the API, as the conversation’s sender.
Does n8n LinkedIn automation raise an account’s daily limit?
No. Each sender starts at 15 connection requests a day, up to 25, and never more than 100 a week, whatever adds the leads. More leads wait their turn. More senders add capacity, each on its own owner’s account.
Which events can our webhooks send to n8n?
connection.sent, connection.accepted, message.sent, message.replied, message.received, lead.tagged, lead.finished and campaign.finished. An endpoint can be limited to chosen events and to one campaign, and carries up to 5 custom headers.