What it does
It can also run without a key, using a non-Jev baseline.
Agent tooling
A LiteLLM router that asks Jev to pick a model for each request.
Only you can see your notes.
Screenshot unavailable. Open the experiment ↗
It can also run without a key, using a non-Jev baseline.
Start by listing the different AI models your app might need and what each one does well. Have a developer set up this tool to direct user messages to the right model automatically. Your app can run the cheapest capable model on its own, or you can supply an access key to let TypeSafe choose.
Remember that letting TypeSafe make the choice sends a short summary of each prompt to that outside service. Do not enable that setting if your project handles private or sensitive user text. If an error occurs, the system switches to a default model you select in advance.