@alloralabs/game-allora-plugin v0.1.2

Allora Plugin for Virtuals Game

The Allora Network plugin seamlessly empowers G.A.M.E agents with real-time, advanced, self-improving AI inferences, delivering high-performance insights without introducing any additional complexity.

Features

• Get price inferences for various assets and timeframes
• Get all available topics on Allora Network
• Fetch inferences by topic ID

Available Functions

1. getPriceInference: Fetches the price inference for the specified asset and a timeframe
2. getAllTopics: Retrieves all available topics on Allora Network
3. getInferenceByTopicId: Fetches the latest inference for a specific topic

Installation

To install the plugin, use npm or yarn:

npm install @alloralabs/game-allora-plugin

or

yarn add @virtuals-protocol/game-allora-plugin

Usage

Importing the Plugin

First, import the AlloraPlugin class from the plugin:

import AlloraPlugin from "@virtuals-protocol/game-allora-plugin";

Setup environment variables

Set the following environment variables:

• ALLORA_API_KEY: Create an API key by creating an account.
• ALLORA_CHAIN_SLUG (Optional): Must be one of: mainnet, testnet. Default value: testnet

Creating a Worker

Create a worker with the necessary Allora API client config:

const alloraPlugin = new AlloraPlugin({
  apiClientConfig: {
    chainSlug: process.env.ALLORA_CHAIN_SLUG as ChainSlug, // Should be one of "testnet" or "mainnet". Default: "testnet"
    apiKey: process.env.ALLORA_API_KEY, // Default key: UP-17f415babba7482cb4b446a1
  },
});

Creating an Agent

Create an agent and add the worker to it:

import { GameAgent } from "@virtuals-protocol/game";

const agent = new GameAgent("GAME_API_KEY", {
  name: "Allora Worker",
  goal: "Retrieve the price of ETH in 5 minutes.",
  description: "You are an AI agent able to fetch price inferences and topic inferences from Allora Network.",
  workers: [alloraPlugin.getWorker({})],
});

Running the Agent

Initialize and run the agent:

(async () => {
  await agent.init();

  while (true) {
    await agent.step({
      verbose: true,
    });
  }
})();