Description

@deepset-ai/n8n-nodes-haystack-enterprise-mcp

This is an n8n community node. It lets you use pipelines you’ve enabled as MCP tools on a
Haystack Enterprise Platform workspace as tools for an
n8n AI Agent.

New to Haystack Enterprise Platform? Start with the
getting started guide.

n8n is a fair-code licensed
workflow automation platform.

Contents

• Installation
• Operations
• Example workflows
• Haystack Enterprise AI companion node
• Credentials
• Resources
• Version history
• Contributing

Installation

This node works on self-hosted n8n only (2.16 or later in the 2.x series; we also check it
against the latest n8n release). It is not n8n-verified, so it can’t run on n8n Cloud.

Enable community nodes on your instance, then follow the
installation guide and enter the
package name @deepset-ai/n8n-nodes-haystack-enterprise-mcp.

Looking for the n8n-verified Haystack Enterprise nodes? See the
haystack-enterprise-ai below.

Operations

The node connects to the MCP server of a Haystack Enterprise Platform workspace and exposes its
enabled pipelines as tools. Add it to the Tools input of an AI Agent, pick your workspace, and
the Agent can call your pipelines.

• Workspace: the workspace whose pipelines to expose.
• Tools to Include: which of the workspace’s tools the Agent can use:

– All: every tool (default).
– Selected: only the tools you pick.
– All Except: every tool except the ones you pick.

• Options → Timeout: how long to wait for a pipeline call, in ms (default 60000).

Example workflows

Try the
AI Agent with MCP tools example:
an AI Agent with this node wired into its Tools input.

To use it, open the file on GitHub, click Copy raw file, and paste it onto the n8n canvas
(Ctrl/Cmd+V). Then add your own credentials (including for the OpenAI Chat Model node) and select
your workspace on the Haystack Enterprise MCP node.

Haystack Enterprise AI companion node

Use this node to give an AI Agent access to your pipelines as tools.

For other Haystack Enterprise tasks in n8n, use the
verified companion package
(source). It works on n8n Cloud and
includes:

• Haystack Enterprise AI — Chat with deployed services, upload and manage files and metadata,

look up indexes, and manage search sessions.

• Haystack Enterprise Chat Model is a language-model sub-node for n8n’s AI Agent that runs on a

Haystack Enterprise service.

Credentials

The node’s Authentication dropdown offers two ways to authenticate.

API Key

Choose API Key on the node (the default) and create a “Haystack Enterprise API” credential:

1. In Haystack Enterprise Platform, go to Settings → API Keys and
generate a key. See Generate an API key.
2. In n8n, paste the key into the credential. Server URL defaults to the hosted Haystack
Enterprise Platform API (https://api.cloud.deepset.ai); for dev or private environments, set it
to your instance’s API origin.

The key needs at least workspace and search permissions.

OAuth2 (sign in with your Haystack Enterprise account)

OAuth2 credentials require n8n 2.19.2 or later.

Choose OAuth2 on the node and create a “Haystack Enterprise OAuth2 API” credential.

1. Leave Server URL at https://api.cloud.deepset.ai. The Server URL is the API origin the
nodes use for requests and OAuth2 discovery; change it only for dev or private Haystack
Enterprise deployments.
2. Click Connect and sign in with your Haystack Enterprise account.

Manual OAuth2 configuration. Turn Use Dynamic Client Registration off to configure or reuse
an existing OAuth2 client, including clients for private Haystack Enterprise deployments. Keep
Server URL pointed at the Haystack Enterprise API, fill in the OAuth2 settings, and register the
callback URL shown on the credential as a redirect URI on the client. Leave Client Secret empty
for a public PKCE client; confidential clients still require their secret.

Resources

• n8n community nodes documentation
• Use your pipelines as MCP tools
• Haystack Enterprise Platform
• Haystack Enterprise Platform docs

Version history

See the changelog.

Contributing

See
CONTRIBUTING.md.
Report security issues as described in
SECURITY.md.