Use Vancine Models in Pi Without Maintaining models.json

Vancine is now available in Pi through the community extension pi-provider-vancine.

Install

pi install npm:pi-provider-vancine

Restart Pi after installation.

Log in

Inside Pi, run:

/login

Choose Vancine, then paste your own Vancine API key into Pi's secret prompt. The extension uses Pi's credential store and does not create a separate credential file.

Choose a model

Run:

/model

Select a model under the Vancine provider and start using it. You do not need to create or maintain ~/.pi/agent/models.json for this setup.

The extension loads compatible Chat Completions models from Vancine's dynamic catalog. It caches and revalidates the catalog, so this is not a claim that every startup fetch is realtime.

Links

• Pi Package Catalog
• npm package
• GitHub source
• Vancine setup guide

pi-provider-vancine is a community extension published and maintained by Vancine. It is not an official Pi or Earendil Works extension, partnership, certification, or endorsement.