Skip to main content

Jama Connect Help

Preparing your model adapter

Your organization operates the adapter that connects Jama Intelligence Gateway to its selected model provider or local inference runtime. The adapter implements the documented Jama Intelligence Gateway interface and translates requests and responses between Jama Connect and the provider.

You can implement the adapter in an existing API, programmable proxy or gateway, serverless function, standalone service, or another platform your organization approves. Jama Software provides the interface documentation and starter kit but doesn’t host the adapter service.

Adapter overview
image4.png

Jama Connect sends requests to the complete operation URL you configure. The adapter translates those requests for the selected provider, then returns responses in the Jama Intelligence Gateway format. Configure the provider credentials, model or deployment identifier, region, and inference settings in the adapter environment. Jama Connect doesn’t supply those values.

A direct provider API URL or an unconfigured pass-through proxy doesn’t implement the Jama Intelligence Gateway interface. The endpoint must translate the request and response formats, including conversation history and Jama Connect tool workflows.

Before you begin

Confirm that the adapter:

  • Is reachable from your Jama Connect environment.

  • Exposes a complete HTTP or HTTPS POST operation URL that accepts application/json.

  • Stores provider credentials securely in the adapter environment.

  • Connects to a model that supports the chat, structured response, and tool behaviors required by your intended Assistant workflows.

  • Uses logging and monitoring that don’t expose credentials or unnecessary customer content.