Anvia Docs

Use @anvia/gemini for Gemini API and Vertex AI.

import { AgentBuilder } from "@anvia/core";
import { GeminiClient } from "@anvia/gemini";

const gemini = new GeminiClient({ apiKey });
const model = gemini.completionModel("gemini-2.5-flash");

const agent = new AgentBuilder("assistant", model)
  .instructions("Answer directly.")
  .build();

Vertex AI uses the same client with explicit project and location values:

const vertex = new GeminiClient({
  vertexai: true,
  project,
  location,
});

Models

Capability	Example
Completion	`gemini.completionModel("gemini-2.5-flash")`
Embeddings	`gemini.embeddingModel("gemini-embedding-001")`
Vertex AI	`new GeminiClient({ vertexai: true, project, location })`

Credentials are passed explicitly to the constructor. Anvia does not read environment variables.

Google Gemini

Models