Introduction

Write TypeScript, define routes, and deploy with a single command. We handle the infrastructure, from automatic scaling to built-in observability and more.

Here's a basic example:

import { createAgent } from '@agentuity/runtime';
import { generateText } from 'ai';
import { openai } from '@ai-sdk/openai';
import { s } from '@agentuity/schema';
 
const agent = createAgent('Chat', {
  schema: {
    input: s.object({ message: s.string() }),
    output: s.object({ response: s.string() }),
  },
  handler: async (ctx, input) => {
    const { text } = await generateText({
      model: openai('gpt-5-mini'),
      prompt: input.message,
    });
    return { response: text };
  },
});
 
export default agent;

This agent uses the AI SDK to call OpenAI and respond to messages. Deploy with agentuity deploy and it scales automatically. Call it from web apps, other agents, or any HTTP client.

What You Get

• Routes & Triggers: HTTP, WebSocket, SSE, cron
• Storage: Key-value, vector, object storage, durable streams
• Sandbox: Execute code in isolated containers with controlled resources and network access
• Evaluations: Automated quality checks that run after each agent execution
• AI Gateway: Route LLM calls through OpenAI, Anthropic, Google, and more
• Type Safety: Built-in schema validation with @agentuity/schema, Zod, or Valibot
• Observability: Logging, tracing, real-time analytics
• App: Visual dashboard for deployments, logs, and more
• Frontend: Deploy React apps alongside your agents

How It Works

Each agent lives in its own file under src/agent/. Routes are defined separately in src/api/index.ts:

Infrastructure like cron schedules is defined in route code, not config. Rolling back a deployment restores the exact configuration from that version.

We're building for a future where agents are the primary way to build and operate software, and where infrastructure is purpose-built for this new paradigm.

Next Steps

• Installation: Set up your environment
• Quickstart: Build your first agent in 5 minutes