Separate virtual machines for langflow and llama running openwebui

[stondino00]

Hello,

We have 2 virtual machines.

1. Is running langflow on ubuntu.
2. The other is running ollama with openwebui.

Is there a way in langflow to use the remote llama and/or openwebui api connection to run the designed chat bots?

We were hoping we could put in the IP of the llama and openwebui server and the api key and have it run models on that remote instance.

Thanks

[dosubot[bot]]

Hi @stondino00! I'm Dosu and I’m helping the langflow team.

Yes, you can configure Langflow to use a remote Ollama (Llama) server and, in general, connect to remote APIs like OpenWebUI.

For Ollama, in the OllamaModel component within Langflow, set the "Base URL" field to the HTTP address and port of your remote Ollama server (for example, http://<remote-ollama-ip>:11434). This allows Langflow to fetch and use models from the remote machine, as long as the network connection is open between the two VMs and the Ollama API is accessible. If you encounter issues with models not appearing, check your network/firewall settings and consider disabling the "tool-mode" setting in the component, as it can filter out models that do not support tool calling [source] [source].

For OpenWebUI or other APIs, you can use Langflow’s APIRequestComponent to make HTTP requests to any remote API endpoint. In this component, you can specify the URL (including the remote server’s IP), HTTP method, headers (such as API keys), and body as needed. This enables integration with external APIs, including OpenWebUI, provided you know the API endpoints and authentication requirements [source].

If you are running Langflow in Docker and Ollama on the host or another VM, ensure you use the correct network address (e.g., http://host.docker.internal:11434 for Docker-to-host communication) [source].

Make sure both VMs can communicate over the network, and that any required API keys are set in the appropriate fields or headers. Langflow manages API keys and their validity, so ensure your keys are correct and valid for the remote services [source].

If you run into issues with the Ollama base URL field being cleared or models not loading, this may be a known issue—try re-entering the URL or checking for updates [source].

In summary: set the remote server’s address in the relevant component’s configuration, provide any necessary API keys, and ensure network connectivity between the VMs. This setup allows Langflow to run chatbots using models hosted on your remote Ollama or OpenWebUI instance.

To reply, just mention my name @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}