The full-stack platform for AI agents. Build with type-safe schemas, frontend hooks, and real-time routes. Ship agents as easily as web apps.

v1 now in public preview — Feedback welcome on Discord and GitHub Discussions.

From local government to indie developers, people are already building with us.

Let's see this in action - a complete flow from agent to API to React frontend:

1. The Agent (src/agent/chat/agent.ts)

import { createAgent } from '@agentuity/runtime';
import { s } from '@agentuity/schema';
import { generateText } from 'ai';
import { openai } from '@ai-sdk/openai';

const agent = createAgent('chat', {
  description: 'A simple chat agent',
  schema: {
    input: s.object({ message: s.string() }),
    output: s.object({ response: s.string() }),
  },
  handler: async (ctx, { message }) => {
    const { text } = await generateText({
      model: openai('gpt-5-mini'),
      prompt: message,
    });
    return { response: text };
  },
});

export default agent;

2. The API Route (src/api/index.ts)

import { createRouter } from '@agentuity/runtime';
import chat from '@agent/chat';

const router = createRouter();

router.post('/chat', chat.validator(), async (c) => {
  const data = c.req.valid('json');
  const result = await chat.run(data);
  return c.json(result);
});

export default router;

3. The React Hook (src/web/components/Chat.tsx)

import { useAPI } from '@agentuity/react';

export function Chat() {
  const { invoke, data, isLoading } = useAPI('POST /api/chat');

  return (
    <div>
      <button onClick={() => invoke({ message: 'Hello!' })}>
        {isLoading ? 'Thinking...' : 'Send'}
      </button>
      {data && <p>{data.response}</p>}
    </div>
  );
}

Use agentuity dev to test locally, then agentuity deploy to production. 🚀

Check out our SDK (runtime, CLI, frontend hooks, server utilities) and docs.

# Install the CLI
curl -fsS https://v1.agentuity.sh | sh

# Create a new project
agentuity create

# Start developing
cd my-project && agentuity dev

# Deploy to the cloud
agentuity deploy

Everything you need to build and ship full-stack AI agents:

Build

• TypeScript-first: Type-safe schemas, autocomplete everywhere, powered by Bun
• Frontend: Deploy web apps alongside your agents with built-in hooks
• Framework-agnostic: Use any AI library (e.g. Vercel AI SDK, Mastra) or bring your own
• Multi-agent: Coordinate agents with type-safe calls between them
• Offline-ready: Start building immediately, no account required
• Agent-friendly CLI: --json output, --explain previews, --dry-run validation, schema discovery

Connect

• AI Gateway: Access all major LLM providers (e.g. OpenAI, Anthropic, Google) with just one API key
• Infrastructure as code: HTTP, cron, email, SMS, WebSocket, and SSE. Rollback-friendly deployments.
• Storage: Key-value, vector, object, durable streams. BYO supported.
• Custom domains: Automated SSL certificates and DNS management
• Deploy anywhere: Public cloud, private cloud, on-prem, or edge

Monitor

• Observability: OpenTelemetry tracing, structured logging, real-time analytics
• Evaluations: Automated quality checks after each agent run
• Container access: SSH and SCP into running deployments

• Documentation: Guides, examples, and reference
• Discord: Join 200+ developers to chat, ask for help, and share what you're building

Our SDK is open source. Contributions welcome! See the repo for guidelines.