Raycast extensions with Meridian
Bring Meridian models into your Raycast workflow. Built-in Raycast AI is hosted by the Raycast team, but third-party extensions accept custom OpenAI-compatible endpoints — point them at Meridian and keep everything under your own keys.
1. Get your endpoint
Copy your Meridian API base URL and key from the API keys dashboard. The base URL follows the OpenAI chat-completions shape:
https://api.meridian.sh/v1
2. Configure the extension
Inside any Raycast extension that supports custom OpenAI providers, set the API base to your Meridian endpoint and paste your key. Model names follow the standard Meridian catalog — use meridian-pro or meridian-fast.
3. Lock it down
Raycast extensions run locally. Your Meridian key never touches Raycast servers when using a third-party extension with a custom endpoint. For additional hardening, scope the key to the extension IP range in your Meridian dashboard.
Built-in Raycast AI is a separate hosted product. This guide covers third-party extensions that let you bring your own provider. Questions? Contact support.