Adding pinstripes as an OpenAI connection in Open WebUI

What you'll need

- [Open WebUI](https://github.com/open-webui/open-webui) running locally or self-hosted

- A pinstripes API key — get one at [pinstripes.io/signup](/signup)

Step 1 — Open admin settings

Log in to Open WebUI as an admin. Go to **Settings → Admin Settings → Connections**.

Step 2 — Add a new OpenAI connection

Click **+ Add connection** (or edit the existing OpenAI entry). Set:

Click **Save**, then **Verify connection**. You should see a green checkmark.

Step 3 — Use pinstripes models

Open a new chat. In the model selector, pinstripes' models will appear alongside any locally-running Ollama models:

- ps/qwen3.6-a3b

- ps/deepseek-v4-flash

- ps/zai-glm-4.5-air

- ps/gpt-oss-120b

Docker deployment note

If Open WebUI is running in Docker and your pinstripes base URL needs to reach the host network, you can use https://api.pinstripes.io/v1 directly — it resolves externally and works from inside any container without host networking tricks.

Using pinstripes + Ollama together

Open WebUI supports multiple model sources simultaneously. You can keep your local Ollama models for offline use while routing larger or production requests to pinstripes. Just add both connections and pick the right model per chat.

Troubleshooting

**Models don't appear**: After saving the connection, refresh the model list using the circular-arrow icon in the model picker.

**"Failed to fetch"**: Your Open WebUI instance may be blocking external requests. Check that https://api.pinstripes.io is not in any blocklist in your Open WebUI environment config.