How do I add an MCP server to n8n?

1. Open the workflow and add an AI Agent node 2. Attach an MCP Client Tool node to it 3. Set the endpoint URL and choose the authentication type 4. Store any API key as an n8n credential rather than in the node 5. Execute the node once to load the tool list

Where is the n8n MCP configuration stored?

There is no file to edit. Servers are added through the interface, and any API key is stored in the platform's own credential store.

Can n8n run a local MCP server?

No. n8n cannot start a local process, so it only works with servers that publish an HTTP or SSE endpoint.

Which parts of the protocol does n8n not support?

It does not implement resources, prompts, sampling or elicitation. A server whose value is mostly in those features will feel incomplete here.

Does n8n also work as an MCP server?

Yes. n8n publishes its own MCP server, so it can be called by other clients as well as calling servers itself.

Does n8n need restarting after adding a server?

No, it picks up the change without a restart.

When was this setup information last checked?

These details were verified on 2026-07-15. Clients change their configuration format from time to time, so check the official documentation if something does not work.

n8n

Workflow automation platform. An MCP Client Tool node calls servers from inside a workflow, and n8n can also expose its own workflows as a server.

Automation platform also publishes a server verified 2026-07-15

How to add a server to n8n

n8n runs on a server, so it cannot start a local stdio process. Only servers with an HTTP or SSE endpoint work here.

Open the workflow and add an AI Agent node
Attach an MCP Client Tool node to it
Set the endpoint URL and choose the authentication type
Store any API key as an n8n credential rather than in the node
Execute the node once to load the tool list

Where the configuration lives

Configured in the interface, not in a file.

Local server configuration

Not supported by this client.

Hosted server configuration

Not supported by this client.

Placeholders such as {{name}}, {{command}}, {{args}} and {{url}} are filled from each server's package or endpoint data.

Protocol support

Support is uneven across clients. A server that only exposes prompts is of no use in a client that does not implement them.

Transports
Transport	Supported
stdio	No
Streamable HTTP	Yes
SSE	Yes

Features
Feature	Supported
Tools	Yes
Resources	No
Prompts	No
Sampling	No
Roots	No
Elicitation	No
OAuth	Yes

Official n8n MCP documentation

Frequently asked questions

How do I add an MCP server to n8n?: 1. Open the workflow and add an AI Agent node 2. Attach an MCP Client Tool node to it 3. Set the endpoint URL and choose the authentication type 4. Store any API key as an n8n credential rather than in the node 5. Execute the node once to load the tool list
Where is the n8n MCP configuration stored?: There is no file to edit. Servers are added through the interface, and any API key is stored in the platform's own credential store.
Can n8n run a local MCP server?: No. n8n cannot start a local process, so it only works with servers that publish an HTTP or SSE endpoint.
Which parts of the protocol does n8n not support?: It does not implement resources, prompts, sampling or elicitation. A server whose value is mostly in those features will feel incomplete here.
Does n8n also work as an MCP server?: Yes. n8n publishes its own MCP server, so it can be called by other clients as well as calling servers itself.
Does n8n need restarting after adding a server?: No, it picks up the change without a restart.
Is there anything specific to watch out for with n8n?: n8n runs on a server, so it cannot start a local stdio process. Only servers with an HTTP or SSE endpoint work here.
When was this setup information last checked?: These details were verified on 2026-07-15. Clients change their configuration format from time to time, so check the official documentation if something does not work.

Other clients

Claude Desktop Claude Code Cursor Visual Studio Code Cline Windsurf