Providers

Google Gemini and Vertex AI

Use google_gemini for Google AI Studio Gemini requests and vertex_ai for Vertex AI Gemini-style deployments.

Gemini generateContent

const { request } = await connector.createProviderContextRequest(
  "google_gemini",
  {
    task: "Use saved project context to prepare a product brief.",
    tokenBudget: 1600
  },
  {
    model: "gemini-3.5-flash",
    input: "Write the brief for the launch review."
  }
);

// Send `request` with your Gemini client.

The SDK maps Neutron memory into config.systemInstruction and maps user input into contents.

Vertex AI

const { request } = await connector.createProviderContextRequest(
  "vertex_ai",
  { task: "Answer with workspace memory and current task details." },
  {
    model: process.env.VERTEX_GEMINI_MODEL,
    input: "Prepare the implementation checklist."
  }
);

Send the generated request through your Vertex AI client or routing layer. Keep Vertex project, region, and service account credentials in your application.