# Open WebUI (/docs/guides/open-webui)

Chat with cheapaimodel models in Open WebUI through an OpenAI API connection.



Open WebUI reaches OpenAI-compatible services through **OpenAI API** connections, which an admin adds under **Settings → Admin → Connections**. Once the connection is saved, its models show up in every user's model selector.

|          | Open WebUI setting                          | Value                                                                             |
| -------- | ------------------------------------------- | --------------------------------------------------------------------------------- |
| Base URL | **URL**                                     | `https://api.cheapaimodel.ai/v1`                                                  |
| API key  | **API Key** (with **Auth** on **Bearer**)   | Your key from the console's [API Keys](https://console.cheapaimodel.ai/keys) page |
| Model ID | Model selector, or **Advanced → Model IDs** | `deepseek-v4-1-flash`, or any ID from the [catalog](/models)                      |

## Set it up [#set-it-up]

<Steps>
  <Step>
    ### Add the connection [#add-the-connection]

    1. Signed in as an admin, open **Settings → Admin → Connections**.
    2. Next to **Manage OpenAI API Connections**, select **Add Connection*&#x2A; (the &#x2A;*+** button).
    3. Set **URL** to `https://api.cheapaimodel.ai/v1`.
    4. Leave **Auth** on **Bearer** and paste your cheapaimodel key into **API Key**.
    5. Leave **API Type** on **Chat Completions**.
    6. Select **Verify Connection**, then **Save**.
  </Step>

  <Step>
    ### Pick a model [#pick-a-model]

    Open WebUI reads the model list from `/v1/models`, which returns the models your key may use. Choose `deepseek-v4-1-flash` in the model selector and start a chat.

    To offer only some models, add their IDs under **Advanced → Model IDs**; Open WebUI then lists exactly those and stops fetching the list. If another connection serves the same IDs, set a **Prefix ID** such as `cheapaimodel`: the model appears as `cheapaimodel.deepseek-v4-1-flash`, and Open WebUI strips the prefix before sending the request.
  </Step>
</Steps>

## Environment variables [#environment-variables]

On a new install you can create the same connection at startup. Add the two variables to the command or Compose file you start Open WebUI with:

```bash title="Terminal"
docker run -d -p 3000:8080 \
  -v open-webui:/app/backend/data \
  -e OPENAI_API_BASE_URL="https://api.cheapaimodel.ai/v1" \
  -e OPENAI_API_KEY="$CHEAPAIMODEL_API_KEY" \
  --name open-webui --restart always \
  ghcr.io/open-webui/open-webui:main
```

Open WebUI reads these only on its first start and then keeps the values in its database. Change the connection later in **Settings → Admin → Connections**.

## Documents and knowledge [#documents-and-knowledge]

cheapaimodel has no embeddings models yet. Leave the embedding engine under **Settings → Admin → Documents** on Open WebUI's built-in model or another provider; don't point it at cheapaimodel. The `:slim` image has no built-in embedding model, so RAG there needs another provider's embeddings.

## Common mistakes [#common-mistakes]

* **Verify Connection fails with `404` and `fault filter abort`, and no models appear.** **URL** is missing `/v1`. Set `https://api.cheapaimodel.ai/v1`.
* **Every message fails with `404` and `Invalid URL (POST /v1/chat/completions/chat/completions)`.** **URL** includes the endpoint path. Open WebUI adds the path itself; set `https://api.cheapaimodel.ai/v1`.
* **`401`.** The key is wrong, expired or out of budget, or **Auth** is not **Bearer**. Check the key in the console.
* **`503` with `model_not_found`.** An ID typed into **Model IDs** doesn't match the [catalog](/models). Fix the spelling, or clear the list to fetch it from `/v1/models`.
* **Changed environment variables have no effect.** They only apply on the first start. Edit the connection in **Settings → Admin → Connections**.