What causes "'[Errno 111] Connection refused'" under Results > Errors?

[carlh04426]

From 'docker container ls -a:
103fceec2c76 web-ui-browser-use-webui "/usr/bin/supervisor…" About an hour ago Up About an hour (healthy) 0.0.0.0:5901->5901/tcp, :::5901->5901/tcp, 0.0.0.0:6080->6080/tcp, :::6080->6080/tcp, 0.0.0.0:7788->7788/tcp, :::7788->7788/tcp, 0.0.0.0:9222->9222/tcp, :::9222->9222/tcp web-ui-browser-use-webui-1

1. Web Interface is accessible at http://localhost:7788 (error message is displayed here under Results > Errors)
2. VNC Interface is accessible at http://localhost:6080/vnc.html
3. The browser has been instantiated and says "Browser Ready"
4. When I click "Run Agent", a new 'about:blank' tab is opened in the browser and the above error appears after several seconds.
5. $ ollama ps
   NAME ID SIZE PROCESSOR UNTIL
   deepseek-r1:7b 0a8c26691023 6.0 GB 100% GPU 4 minutes from now

[zhampu]

same here

[jes83]

same for me. Can't connect to Ollama server hosted as container too.

[AliYmn]

@zhampu @jes83 @carlh04426
If you're working with Ollama, you should update the OLLAMA_ENDPOINT to "http://host.docker.internal:11434" in your .env file. This allows the Docker container to communicate with the Ollama application running on your host machine, as "localhost" inside a container refers to the container itself, not your computer.

also I made pr about this issue : #399

[carlh04426]

Hi @AliYmn :)

I suspected something like this, thank you!

[carlh04426]

@AliYmn Addendum: Updating the OLLAMA_ENDPOINT to "http://host.docker.internal:11434/" in your .env file may work on Windows or Mac but it doesn't work on Linux. Here's all it took on my Linux system (with native docker installed since well before docker-ce / docker desktop etc.):

• • • • • 8< - - - - -
          network_mode: "host" # <<== *** inserted this line into docker-compose.yml ***
          ports:
      • "7788:7788" # Gradio default port
      • "6080:6080" # noVNC web interface
      • "5901:5901" # VNC port
      • "9222:9222" # Chrome remote debugging port
• • • • • 8< - - - - -
          Success! (sort of) :) However this application seems to require more compute than my seven year old i7 laptop w/ a GeForce GTX 1070 Mobile 8GB (GDDR5 VRAM) can deliver. Ollama thought for ten minutes while generating a system load factor of ~5.86 yielding no visible changes in the VNC viewer, so I killed the container.

[AliYmn]

@carlh04426 For Linux we can try to use the ip address of the device, can you try this method? (I have added a few alternative solutions.)

1. Use Host IP Address

# Find your host IP (usually 172.17.0.1 for Docker)
ip -4 addr show docker0 | grep -Po 'inet \K[\d.]+'

# Then use it in your config
OLLAMA_ENDPOINT=http://172.17.0.1:11434

2. Add Host Definition

In docker-compose.yml:

extra_hosts:
  - "host.docker.internal:host-gateway"

Then use OLLAMA_ENDPOINT=http://host.docker.internal:11434

3. Use Host Network

In docker-compose.yml:

network_mode: "host"

Then use OLLAMA_ENDPOINT=http://localhost:11434

4. Use Ngrok as a Last Resort

If none of the above solutions work, you can use Ngrok to create a tunnel:

# Install and run Ngrok
ngrok http 11434

# Use the provided URL in your config
OLLAMA_ENDPOINT=http://your-ngrok-url

This creates a public URL forwarding to your local Ollama instance, which works regardless of network configuration.

[int75]

I have similar challenge.

OS: ollama in MacOS

NAME          ID              SIZE     PROCESSOR    UNTIL              
qwq:latest    009cb3f08d74    23 GB    100% GPU     4 minutes from now 

web-ui for ollama works fine at http://locahost:3000

OLLAMA_ENDPOINT in .env
OLLAMA_ENDPOINT=http://host.docker.internal:11434/

web-ui for ollama LLM configuration

Error in terminal
To create a public link, set share=True in launch().
INFO [agent] 🚀 Starting task: go to google.com and type 'OpenAI' click search and give me the first url
INFO [src.agent.custom_agent]
📍 Step 1
ERROR [agent] ❌ Result failed 1/3 times:
[Errno 8] nodename nor servname provided, or not known
INFO [src.agent.custom_agent]
📍 Step 1
ERROR [agent] ❌ Result failed 2/3 times:
[Errno 8] nodename nor servname provided, or not known
INFO [src.agent.custom_agent]
📍 Step 1
ERROR [agent] ❌ Result failed 3/3 times:
[Errno 8] nodename nor servname provided, or not known
ERROR [agent] ❌ Stopping due to 3 consecutive failures
INFO [src.agent.custom_agent] Created GIF at agent_history.gif

What am I doing wrong or have not done?

[manwithsteelnerves]

I have similar challenge.

OS: ollama in MacOS

NAME          ID              SIZE     PROCESSOR    UNTIL              
qwq:latest    009cb3f08d74    23 GB    100% GPU     4 minutes from now 

web-ui for ollama works fine at http://locahost:3000

OLLAMA_ENDPOINT in .env OLLAMA_ENDPOINT=http://host.docker.internal:11434/

web-ui for ollama LLM configuration

Error in terminal To create a public link, set share=True in launch(). INFO [agent] 🚀 Starting task: go to google.com and type 'OpenAI' click search and give me the first url INFO [src.agent.custom_agent] 📍 Step 1 ERROR [agent] ❌ Result failed 1/3 times: [Errno 8] nodename nor servname provided, or not known INFO [src.agent.custom_agent] 📍 Step 1 ERROR [agent] ❌ Result failed 2/3 times: [Errno 8] nodename nor servname provided, or not known INFO [src.agent.custom_agent] 📍 Step 1 ERROR [agent] ❌ Result failed 3/3 times: [Errno 8] nodename nor servname provided, or not known ERROR [agent] ❌ Stopping due to 3 consecutive failures INFO [src.agent.custom_agent] Created GIF at agent_history.gif

What am I doing wrong or have not done?

I see the same exact issue.
Can someone help on this?