Unable to Connect n8n to Local Ollama Instance (ECONNREFUSED Error)

Questions
1

Describe the problem/error/question

I am trying to integrate n8n with the Ollama model installed locally on my PC. However, I am unable to establish a connection. I have tried configuring the Base URL with the following values:

  • http://localhost:11434/
  • http://127.0.0.1:11434/

Despite multiple attempts, the connection fails, displaying an ECONNREFUSED error.

I have searched the documentation, YouTube tutorials, and online resources, but none of them mention encountering this issue. Most guides simply use one of the above URLs, and Ollama connects successfully.

I would appreciate any guidance on resolving this issue.

What is the error message (if any)?

Couldnā€™t connect with these settings ECONNREFUSED

Image
Image1328Ɨ670 32.4 KB

Image
Image1346Ɨ688 32.6 KB

Please share your workflow/screenshots/recording

(Screenshots attached)

Screenshot 1
Screenshot 2

Share the output returned by the last node

N/A (The issue occurs during the connection setup, preventing workflow execution.)

#[details=ā€œinstance informationā€]

Debug info

core

  • n8nVersion: 1.82.3
  • Platform: Windows 10 (Local Installation)
  • Node.js Version: 20.18.3
  • Database: SQLite
  • Execution Mode: Regular
  • Concurrency: 5
  • License: Community

storage

  • Success logs: All
  • Error logs: All
  • Progress tracking: Disabled
  • Manual executions: Enabled
  • Binary Mode: Filesystem

pruning

  • Enabled: Yes
  • Max Age: 168 hours
  • Max Count: 2500 executions

client

  • User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/134.0.0.0 Safari/537.36
  • Is Touch Device: No

Generated at: 2025-03-18T12:07:41.379Z
[/details]

1 Like
2

@Sahil_Kayastha can you give a little more detail about your Ollame installation? Is the GUI http://localhost:3000 accessible? Is there a local firewall blocking the port?
Do you know the this docu: ollama common-issues

3

I have the same issue. @Sahil_Kayastha did you manage to find a solution?

In my case Iā€™m running Ollama locally (not in a container) and my n8n workflow is hosted in the n8n cloud. Both http://localhost:11434/ and http://127.0.0.1:11434/ return 200 status and Ollama is running text. According to the documentation this should be enough but I am also getting ECONNREFUSED

4

I still get the same and I have tried [localhost]/api/tags on an n8n HTTP Request node which results in the same error, whereas a curl or call from the browser is OK. This means the issue is ā€œsimplyā€ that localhost wonā€™t accept external connections.

Setting up n8n locally resolves the issue although means importing or recreating your workflow locally - Iā€™m a novice on this platform so maybe thereā€™s an easy way to transfer workflows between the cloud and local.

To set up n8n locally:

npm install -g n8n
n8n start

Allowing connections from the n8n cloud to Ollama localhost could also work but requires managing security issues, so Iā€™m NOT recommending it.

5

If you want ollama to listen on all network devices on your machine then youā€™ll need to set the following environment variable:

export OLLAMA_HOST="0.0.0.0"

Then just start ollama with: ollama serve.

There is a nice FAQ how to run ollama as a server instance: ollama/docs/faq.md at main Ā· ollama/ollama Ā· GitHub

Also if your ollama server is behind a firewall (NAT etc) youā€™ll need to portforward the port from your firewall/router to your machine.

1 Like
6

turn off your ipv6 ,set url as http://127.0.0.1(or real ip):11434,than restart n8n.

7

connect with this:
http://host.docker.internal:11434
I had the same issue when running the ollama models locally but not in a container

5 Likes
8

@Sahil_Kayastha were you able to find a solution?

9

You have this error because n8n inside his own container cannot see the model in ollama.

Solution 1:

Create a network

docker network create n8n-network

run Ollama on this network

docker run -d --name ollama --network n8n-network -p 11434:11434 ollama/ollama

run n8n on the same network

docker run -it --rm --name n8n --network n8n-network -p 5678:5678 -v n8n_data:/home/node/.n8n docker.n8n.io/n8nio/n8n

In ollama use :

http://ollama:11434

Solution 2:

You can also use a docker compose file docker-compose.yml to do the same things

version: ā€˜3.8ā€™
services:
ollama:
image: ollama/ollama
container_name: ollama
ports:
- ā€œ11434:11434ā€
volumes:
- ollama_data:/root/.ollama

n8n:
image: docker.n8n.io/n8nio/n8n
container_name: n8n
ports:
- ā€œ5678:5678ā€
volumes:
- n8n_data:/home/node/.n8n
depends_on:
- ollama
environment:
- N8N_HOST=0.0.0.0

volumes:
ollama_data:
n8n_data:

Use http://ollama:11434

And Thatā€™s it !!!