> ## Documentation Index
> Fetch the complete documentation index at: https://docs.gcore.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Chat with a model via Open WebUI

[Open WebUI](https://openwebui.com) is an open-source, browser-based chat interface for large language models. Gcore hosts it as a bundled module alongside the model at a separate HTTPS endpoint — enable it during deployment from the Application Catalog.

Open WebUI runs as a separate pod alongside the model pod, so the deployment requires `Inference instance count` quota ≥ 2.

## Deploy a model with Open WebUI enabled

1. In the [Gcore Customer Portal](https://portal.gcore.com/), navigate to **Everywhere Inference > Deployments**.

2. Click **Deploy application from catalog**.

3. Under **Deployment Configuration**, select an application from the catalog. The steps below use **meta-llama/Llama-3.2-1B-Instruct**.

4. Under **Routing placement**, select a region.

5. Under **Application modules**, check the **Open WebUI** checkbox.

   Open WebUI adds a CPU pod to the deployment. The default flavor is **4 vCPU / 16 GiB RAM**.

   <Frame>
     <img src="https://mintcdn.com/gcore/uiRa_jFs2CEr69p9/images/docs/edge-ai/everywhere-inference/ai-models/chat-with-open-webui/webui-1.png?fit=max&auto=format&n=uiRa_jFs2CEr69p9&q=85&s=6d58025e89d97fee50dd0b7dbe319db8" alt="Deployment form with Open WebUI module enabled" width="1896" height="836" data-path="images/docs/edge-ai/everywhere-inference/ai-models/chat-with-open-webui/webui-1.png" />
   </Frame>

6. Under **Deployment details**, enter a name for the deployment.

   <Info>
     Deployment names must be alphanumeric. Hyphens are not allowed.
   </Info>

7. Click **Deploy model**.

Gcore provisions both the model pod and the Open WebUI pod. The deployment list shows **Partially deployed** until both pods are ready, then switches to **Active**.

## Open WebUI URL

Once the deployment is **Active**, navigate to **Everywhere Inference > Deployments** and click the deployment name. On the **Overview** tab, the **Endpoints** section shows two URLs:

* `model-<name>-...ai.gcore.dev` — the OpenAI-compatible inference endpoint.
* `ui-<name>-...ai.gcore.dev` — the Open WebUI interface.

Copy the `ui-` URL and open it in a browser.

<Frame>
  <img src="https://mintcdn.com/gcore/uiRa_jFs2CEr69p9/images/docs/edge-ai/everywhere-inference/ai-models/chat-with-open-webui/webui-2.png?fit=max&auto=format&n=uiRa_jFs2CEr69p9&q=85&s=c845b962f492cfa2181f8b9a802509e9" alt="Deployment overview showing both endpoints" width="1249" height="373" data-path="images/docs/edge-ai/everywhere-inference/ai-models/chat-with-open-webui/webui-2.png" />
</Frame>

## Create an admin account

On first access, Open WebUI displays a sign-up form. The first registered user automatically becomes the administrator.

1. On the sign-in page, enter an email and a password.
2. Submit the form. Open WebUI creates the account and logs in immediately.

<Info>
  The admin account is local to this Open WebUI instance and is not linked to the Gcore account.
</Info>

## Start a conversation

After signing in, Open WebUI opens the main chat interface with the deployed model pre-selected.

<Frame>
  <img src="https://mintcdn.com/gcore/uiRa_jFs2CEr69p9/images/docs/edge-ai/everywhere-inference/ai-models/chat-with-open-webui/webui-4.png?fit=max&auto=format&n=uiRa_jFs2CEr69p9&q=85&s=4ddfa3bf7b60fed6021846ef4fd78494" alt="Open WebUI chat interface with model loaded" width="1400" height="900" data-path="images/docs/edge-ai/everywhere-inference/ai-models/chat-with-open-webui/webui-4.png" />
</Frame>

Type a message in the input box and press **Enter** to send it. The model responds in the chat window.

The `model-` endpoint accepts [OpenAI-compatible API requests](/edge-ai/everywhere-inference/ai-models/query-deployed-model) via curl, Python, or JavaScript. To restrict access, enable **API Key authentication** on the deployment **Settings** tab.