Getting Started

Learn the basics of Tora AI and set up your first secure agent

Back to Documentation

Installation Guide

Prerequisites

  • Node.js 18.0 or higher
  • npm or yarn package manager
  • A Tora AI account with API key

Install the SDK

npm install @tora/sdk

Or if you prefer yarn:

yarn add @tora/sdk

Initialize the Client

import { ToraClient } from '@tora/sdk';

// Initialize the client
const client = new ToraClient({
  apiKey: process.env.TORA_API_KEY,
  environment: 'production' // or 'development'
});

// Test the connection
async function testConnection() {
  try {
    const status = await client.system.status();
    console.log('Connection successful:', status);
  } catch (error) {
    console.error('Connection failed:', error);
  }
}

testConnection();

Core Concepts

Agent Security

Tora AI provides a security layer for autonomous AI agents, protecting them from unauthorized access, manipulation, and other threats.

Behavioral Policies

Define rules and constraints for your AI agents to ensure they operate within acceptable parameters and respond appropriately to threats.

Monitoring & Alerts

Real-time monitoring of agent behavior with automated alerts for suspicious activities or policy violations.

Quantum-Resistant Security

Advanced cryptographic techniques designed to withstand attacks from quantum computers, ensuring long-term security.

Quick Start Tutorial

1. Register an Agent

// Register a new agent
const agent = await client.agents.register({
  id: 'agent-123',
  name: 'Assistant Bot',
  capabilities: ['text-generation', 'code-analysis'],
  description: 'A helpful assistant that can generate text and analyze code'
});

console.log('Agent registered:', agent.id);

2. Create a Security Policy

// Create a security policy for the agent
const policy = await client.policies.create({
  agentId: agent.id,
  name: 'Basic Security Policy',
  rules: [
    {
      type: 'rate-limit',
      params: {
        maxRequests: 100,
        timeWindow: '1h'
      }
    },
    {
      type: 'content-filter',
      params: {
        blockedTopics: ['malware', 'hacking'],
        sensitivity: 'medium'
      }
    }
  ]
});

console.log('Policy created:', policy.id);

3. Monitor Agent Activity

// Set up a webhook to receive agent activity events
await client.webhooks.create({
  url: 'https://your-server.com/webhooks/tora',
  events: ['agent.activity', 'security.violation'],
  secret: 'your-webhook-secret'
});

// Or monitor events programmatically
client.events.subscribe('agent.activity', (event) => {
  console.log('Agent activity:', event);
  
  if (event.type === 'suspicious') {
    // Take action on suspicious activity
    client.agents.restrict({
      id: event.agentId,
      reason: 'Suspicious activity detected',
      duration: '1h'
    });
  }
});