Connect WarmySender to n8n

Which route do I need?

A typical setup: a form or CRM trigger feeds new leads into a campaign through an HTTP Request node, a Webhook node catches every reply and posts it to Slack or your CRM, and an AI Agent with the MCP Client Tool answers "which campaigns sent nothing yesterday?" on a schedule.

First, create a WarmySender API key

Two of the three routes need one. In WarmySender go to Settings → API Keys → Create Key, name it (say, n8n) and tick the permissions your workflow needs. The key starts with ws_ and is shown once — copy it into n8n's credential store straight away rather than into a workflow field, so it is encrypted at rest and never appears in an execution log.

Start narrow. If the agent should only report, grant the read permissions. Widen to the write permissions when you actually want it to create prospects or launch campaigns. You can create several keys with different permissions and delete any of them at any time.

Route 1: let an n8n AI Agent run your outreach (MCP Client Tool)

This is the route for "I want to say what I want and have it happen". n8n's AI Agent node can use tools, and WarmySender publishes its whole tool set — campaigns, prospects, warmup, LinkedIn, Instagram, lead search, email verification — as an MCP server, the same open standard Claude, ChatGPT, Cursor and Codex use.

1. Add an AI Agent node to your workflow and attach any chat model you already use in n8n.
2. Add an MCP Client Tool node and connect it to the agent's Tool input.
3. Point it at WarmySender. Set Endpoint to https://warmysender.com/mcp and make sure Server Transport reads HTTP Streamable. That is the default on current n8n releases; on older ones the default was Server Sent Events (Deprecated), and with that setting n8n reports "Could not connect to your MCP server" — switch it and it connects.
4. Authenticate with your key. Set Authentication to Bearer Auth, create a new credential, and paste the ws_ key as the token. (Header Auth also works: header name Authorization, value Bearer ws_….)
5. Choose which tools the agent may use. All is fine to start. To keep an agent on rails, pick Selected and tick only the tools you want it to have — the list loads straight from WarmySender once the credential is valid, which is also your connection test.
6. Try it. Send the agent a chat message or a scheduled prompt: "List my campaigns and tell me which ones sent nothing yesterday." If it answers with your campaign names, you are live.

Each tool the agent can see explains, in its own description, exactly what it does and what it will never do — so a model reading the list learns the rules without you writing a system prompt. The full catalogue, with every tool's description, is on the MCP server page. Prefer signing in over pasting a key? The address also supports the standard sign-in flow, so if your n8n version offers OAuth2 on this node you can approve with your WarmySender account instead.

Route 2: call WarmySender from any workflow (HTTP Request)

For work that should happen exactly the same way every time — a new signup lands in a campaign, a deal closes and the contact is unenrolled, a nightly stats pull into a sheet — skip the model and call the REST API directly.

1. Add an HTTP Request node.
2. Authentication: choose Generic Credential Type → Header Auth. Name: Authorization. Value: Bearer ws_… (your key). Save it as a credential so every WarmySender node in your account can reuse it.
3. Method and URL come from the call you need. The most common one, saving a new lead and enrolling them in a campaign in a single request, is POST https://warmysender.com/api/v1/prospects.
4. Body: turn on Send Body, set Body Content Type to JSON, and map your trigger's fields. For example:

   {
     "email": "{{ $json.email }}",
     "firstName": "{{ $json.firstName }}",
     "company": "{{ $json.company }}",
     "enrollInCampaignId": "YOUR_CAMPAIGN_ID"
   }

   That one call creates the prospect (or merges into an existing one with the same address) and enrolls them; the key needs the prospect-write and campaign-enrollment permissions. Copy the campaign id from the campaign's page in WarmySender.
5. Run it once from the editor and look at the output: a new prospect comes back with its id, and the campaign's enrollment count goes up. From then on it runs on every trigger.

Every call — campaigns, prospects, lists, suppression, warmup, mailboxes, LinkedIn, email verification — is listed with its fields and example responses in the API reference, and the same reference is published as an OpenAPI file at https://warmysender.com/api/v1/openapi.json for any tool that imports one. A ready-made n8n workflow that creates and enrolls a lead, then unenrolls them when they reply on another channel, is walked through there as well.

Route 3: get replies, bounces and accepted invites into n8n (Webhook)

This is the direction most people set up first, because it is where n8n shines: a reply lands and, seconds later, it is in Slack, on a Notion board, in your CRM, or routed to whoever owns that campaign.

1. Add a Webhook node as the trigger of a new workflow. Set HTTP Method to POST and Respond to Immediately, so n8n acknowledges the event straight away rather than after your whole workflow finishes.
2. Publish the workflow and copy the Production URL. The Test URL only listens while you have the editor open, which is why a webhook that "worked yesterday" sometimes stops — it was the test URL.
3. In WarmySender go to Settings → Webhooks → New webhook, paste the URL, tick the events you want, and save. You will be shown a signing secret once.
4. Click Test webhook on the same screen. It arrives in n8n as a webhook.test event so you can map fields before anything real comes through.
5. Chain your action after the trigger — a Slack node with {{ $json.data.prospectEmail }} in the message, an IF node that routes by campaign, a CRM node that logs the activity.

Events you can subscribe to: reply.received, email.opened, email.clicked, email.bounced, email.unsubscribed, prospect.suppressed, limit.hit, linkedin.reply_received (includes the reply text), linkedin.invite_accepted, instagram.reply_received and whatsapp.reply_received. Each payload is signed; if your n8n URL is reachable from the public internet, add a small Code node after the trigger that checks the signature header against your secret before doing anything with the data. The signing scheme and full payload of every event are in the webhooks reference.

What about Make and Zapier?

Exactly the same three shapes, with different node names. There is no branded WarmySender app in either marketplace; you use their generic modules, which is also why nothing here can go stale when they rename a screen.

Both vendors change their AI-agent features often, so for the agent row trust the screen in front of you. The REST and webhook rows have not changed in years and will not.

Frequently asked questions

Can I connect WarmySender to n8n?

Yes, three ways and with no plugin to install. An n8n AI Agent can drive your whole workspace through the MCP Client Tool node pointed at https://warmysender.com/mcp; any workflow can call the REST API from an HTTP Request node with your API key; and a Webhook node can receive every reply, bounce and accepted LinkedIn invite as it happens. It works on self-hosted n8n and on n8n Cloud.

Is there a WarmySender node in n8n?

No, and you do not need one. n8n's built-in HTTP Request, Webhook and MCP Client Tool nodes cover everything, and because they are generic they keep working across n8n upgrades. The same is true of Make and Zapier: their generic HTTP and webhook modules are the integration.

Can an n8n agent send my cold emails?

No agent ever sends anything, and that is deliberate. Your n8n agent can build the campaign, enroll people and start it — and starting it simply hands it to WarmySender's scheduler, which paces every send inside safe daily caps and a gradual ramp. It cannot raise a limit, and it cannot connect or disconnect a mailbox or a social account; that always stays with you in the app. So a runaway workflow cannot burn a domain.

Does it work on my plan?

API keys, the MCP server and webhooks are included with every paid WarmySender plan, including every AppSumo lifetime tier, at no extra cost. The Free plan does not include them. On the n8n side, the AI Agent and MCP Client Tool nodes are part of n8n itself, so they work on a self-hosted instance and on n8n Cloud; you only need a chat model credential for the agent route.

My n8n webhook worked once and then stopped. Why?

Almost always the Test URL. n8n's Test URL only listens while the editor is open and you have clicked to listen; the Production URL is the one that stays registered once the workflow is published. Replace the URL in WarmySender's Settings → Webhooks with the Production URL, then use the Test webhook button to confirm.

The MCP Client Tool node lists no WarmySender tools. What now?

Three things cause nearly every case. The endpoint is not exactly https://warmysender.com/mcp (no trailing path). The credential is not a Bearer token starting with ws_, or it was pasted with a stray space. Or the key carries only some permissions, in which case only the tools it is allowed to reach are listed; create a key with the permissions you need in Settings → API Keys and reselect the credential. And check Server Transport: it must read HTTP Streamable, not the deprecated server-sent-events option — that setting is the one that produces "Could not connect to your MCP server".

Related reading

• API reference — every call, every webhook payload, and the full n8n create-and-enroll workflow
• The WarmySender MCP server — the complete tool reference the MCP Client Tool node loads
• Which AI agents can run outreach — the per-app comparison table, n8n included
• What your AI agent can and cannot do with cold email — the capability matrix
• Connect HubSpot to WarmySender — the CRM version of the same three routes