Transport & Deployment Examples

Express REST Server

Serve any agent over HTTP with createAgentRouter. It auto-generates /agents/:name/run and /agents/:name/stream endpoints for every agent passed in.

import express from "express";
import { Agent, openai, defineTool } from "@agentium/core";
import { createAgentRouter } from "@agentium/transport";
import { z } from "zod";

const lookupOrder = defineTool({
  name: "lookup_order",
  description: "Look up an order by ID",
  parameters: z.object({ orderId: z.string() }),
  execute: async ({ orderId }) => `Order ${orderId}: shipped, arriving Friday`,
});

const agent = new Agent({
  name: "support",
  model: openai("gpt-4o"),
  instructions: "You are a helpful customer-support agent.",
  tools: [lookupOrder],
});

const app = express();
app.use(express.json());
app.use("/api", createAgentRouter({ agents: { support: agent } }));

app.listen(3000, () => console.log("Listening on :3000"));
// POST /api/agents/support/run  { "input": "Where is order 42?" }

Streaming Responses

The router exposes an SSE streaming endpoint at /agents/:name/stream automatically. Clients receive incremental data: frames as the model generates tokens.

import express from "express";
import { Agent, openai } from "@agentium/core";
import { createAgentRouter } from "@agentium/transport";

const agent = new Agent({
  name: "writer",
  model: openai("gpt-4o"),
  instructions: "Write creative short stories.",
});

const app = express();
app.use(express.json());
app.use("/api", createAgentRouter({ agents: { writer: agent } }));
app.listen(3000);

// Client-side consumption:
// const es = new EventSource("/api/agents/writer/stream");
// es.onmessage = (e) => {
//   if (e.data === "[DONE]") return es.close();
//   const chunk = JSON.parse(e.data);
//   process.stdout.write(chunk.text ?? "");
// };

Or call the streaming endpoint directly with fetch:

const res = await fetch("http://localhost:3000/api/agents/writer/stream", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({ input: "Tell me a story about a robot." }),
});

const reader = res.body!.getReader();
const decoder = new TextDecoder();

while (true) {
  const { done, value } = await reader.read();
  if (done) break;
  const text = decoder.decode(value);
  for (const line of text.split("\n")) {
    if (line.startsWith("data: ") && line !== "data: [DONE]") {
      const chunk = JSON.parse(line.slice(6));
      process.stdout.write(chunk.text ?? "");
    }
  }
}

Socket.IO Gateway

For real-time bidirectional communication, use createAgentGateway with Socket.IO. Agents registered in the global registry are immediately available.

import http from "node:http";
import { Server as SocketIO } from "socket.io";
import { Agent, openai } from "@agentium/core";
import { createAgentGateway } from "@agentium/transport";

const server = http.createServer();
const io = new SocketIO(server, { cors: { origin: "*" } });

createAgentGateway({ io });

new Agent({
  name: "assistant",
  model: openai("gpt-4o"),
  instructions: "You are a helpful assistant.",
});

server.listen(3001, () => console.log("WebSocket gateway on :3001"));

// Client:
// const socket = io("http://localhost:3001");
// socket.emit("agent:run", { agent: "assistant", input: "Hello!" });
// socket.on("agent:response", (data) => console.log(data));
// socket.on("agent:stream", (chunk) => process.stdout.write(chunk.text));

Swagger UI

Enable auto-generated OpenAPI docs with interactive Swagger UI by passing swagger config. The spec is derived from your registered agents, teams, and workflows.

import express from "express";
import { Agent, openai } from "@agentium/core";
import { createAgentRouter } from "@agentium/transport";

const agent = new Agent({
  name: "analyst",
  model: openai("gpt-4o"),
  instructions: "You are a data analyst.",
});

const app = express();
app.use(express.json());

app.use(
  "/api",
  createAgentRouter({
    agents: { analyst: agent },
    swagger: {
      enabled: true,
      title: "Agent API",
      description: "AI agent endpoints",
      version: "1.0.0",
      docsPath: "/docs",
      specPath: "/docs/spec.json",
      servers: [{ url: "http://localhost:3000/api", description: "Local" }],
    },
  }),
);

app.listen(3000, () => {
  console.log("API:     http://localhost:3000/api");
  console.log("Swagger: http://localhost:3000/api/docs");
});

File Upload

Accept multimodal inputs (images, PDFs, audio) alongside text by enabling fileUpload. Files are automatically converted into multi-part content arrays.

import express from "express";
import { Agent, openai } from "@agentium/core";
import { createAgentRouter } from "@agentium/transport";

const agent = new Agent({
  name: "vision",
  model: openai("gpt-4o"),
  instructions: "Describe uploaded images in detail.",
});

const app = express();
app.use(express.json());

app.use(
  "/api",
  createAgentRouter({
    agents: { vision: agent },
    fileUpload: {
      maxFileSize: 10 * 1024 * 1024, // 10 MB
      allowedMimeTypes: ["image/png", "image/jpeg", "application/pdf"],
    },
  }),
);

app.listen(3000);
// curl -X POST http://localhost:3000/api/agents/vision/run \
//   -F 'input=What is in this image?' \
//   -F 'file=@photo.jpg'

Per-Request API Keys

In multi-tenant deployments, each request can supply its own API key via headers. The router extracts provider-specific keys and forwards them to the model.

import express from "express";
import { Agent, openai } from "@agentium/core";
import { createAgentRouter } from "@agentium/transport";

const agent = new Agent({
  name: "assistant",
  model: openai("gpt-4o"),
  instructions: "You are a helpful assistant.",
});

const app = express();
app.use(express.json());
app.use("/api", createAgentRouter({ agents: { assistant: agent } }));
app.listen(3000);

// Clients forward their own OpenAI key:
// POST /api/agents/assistant/run
// Headers:
//   X-OpenAI-API-Key: sk-tenant-abc123...
//   Content-Type: application/json
// Body: { "input": "Hello!" }
//
// Supported headers:
//   X-OpenAI-API-Key    → OpenAI
//   X-Google-API-Key    → Google
//   X-Anthropic-API-Key → Anthropic
//   X-API-Key           → Generic fallback

Multiple Agents on One Server

Serve several agents under distinct routes from a single Express app. Each agent gets its own /run and /stream endpoints.

import express from "express";
import { Agent, openai, google, defineTool } from "@agentium/core";
import { createAgentRouter } from "@agentium/transport";
import { z } from "zod";

const getWeather = defineTool({
  name: "get_weather",
  description: "Get weather for a city",
  parameters: z.object({ city: z.string() }),
  execute: async ({ city }) => `${city}: 72°F, sunny`,
});

const weatherBot = new Agent({
  name: "weather",
  model: openai("gpt-4o-mini"),
  instructions: "You provide weather forecasts.",
  tools: [getWeather],
});

const codeBot = new Agent({
  name: "coder",
  model: openai("gpt-4o"),
  instructions: "You are an expert TypeScript programmer.",
});

const researchBot = new Agent({
  name: "researcher",
  model: google("gemini-2.0-flash"),
  instructions: "You research topics and provide summaries.",
});

const app = express();
app.use(express.json());

app.use(
  "/api",
  createAgentRouter({
    agents: {
      weather: weatherBot,
      coder: codeBot,
      researcher: researchBot,
    },
    swagger: { enabled: true, title: "Multi-Agent API" },
  }),
);

app.listen(3000);
// POST /api/agents/weather/run
// POST /api/agents/coder/run
// POST /api/agents/researcher/run
// GET  /api/agents          → list all agents

Teams and Workflows via REST

Serve teams and workflows alongside agents. Each gets typed endpoints under /teams/:name/run and /workflows/:name/run.

import express from "express";
import { Agent, Team, TeamMode, Workflow, openai } from "@agentium/core";
import { createAgentRouter } from "@agentium/transport";

const researcher = new Agent({
  name: "researcher",
  model: openai("gpt-4o-mini"),
  instructions: "Research the given topic.",
});

const writer = new Agent({
  name: "writer",
  model: openai("gpt-4o"),
  instructions: "Write a polished article from research notes.",
});

const team = new Team({
  name: "content-team",
  agents: [researcher, writer],
  mode: TeamMode.SEQUENTIAL,
  instructions: "Research a topic, then write an article about it.",
});

const workflow = new Workflow({
  name: "publish-pipeline",
  steps: [
    { type: "agent", agent: researcher, input: "Research AI trends" },
    { type: "agent", agent: writer },
  ],
});

const app = express();
app.use(express.json());

app.use(
  "/api",
  createAgentRouter({
    agents: { researcher, writer },
    teams: { "content-team": team },
    workflows: { "publish-pipeline": workflow },
    swagger: { enabled: true, title: "Content Platform API" },
  }),
);

app.listen(3000);
// POST /api/teams/content-team/run       { "input": "Write about quantum computing" }
// POST /api/teams/content-team/stream    (SSE)
// POST /api/workflows/publish-pipeline/run

CORS and Security

Configure CORS origins, rate limiting, and custom auth middleware through createAgentRouter options.

import express from "express";
import { Agent, openai } from "@agentium/core";
import { createAgentRouter } from "@agentium/transport";

const agent = new Agent({
  name: "secure-bot",
  model: openai("gpt-4o"),
  instructions: "You are a secure assistant.",
});

function authMiddleware(req: any, res: any, next: any) {
  const token = req.headers.authorization?.replace("Bearer ", "");
  if (!token || token !== process.env.API_TOKEN) {
    return res.status(401).json({ error: "Unauthorized" });
  }
  next();
}

const app = express();
app.use(express.json());

app.use(
  "/api",
  createAgentRouter({
    agents: { "secure-bot": agent },
    cors: ["https://app.example.com", "https://admin.example.com"],
    rateLimit: { windowMs: 60_000, max: 50 },
    middleware: [authMiddleware],
  }),
);

app.listen(3000);

Health Check Endpoint

Add a health and readiness probe alongside the agent router for production deployments behind load balancers.

import express from "express";
import { Agent, openai, registry } from "@agentium/core";
import { createAgentRouter } from "@agentium/transport";

const agent = new Agent({
  name: "bot",
  model: openai("gpt-4o"),
  instructions: "You are a helpful bot.",
});

const app = express();
app.use(express.json());

app.get("/health", (_req, res) => {
  res.json({ status: "ok", timestamp: new Date().toISOString() });
});

app.get("/ready", (_req, res) => {
  const agents = registry.list().agents;
  if (agents.length === 0) {
    return res.status(503).json({ status: "not_ready", reason: "No agents registered" });
  }
  res.json({ status: "ready", agents });
});

app.use("/api", createAgentRouter({ cors: true }));

app.listen(3000, () => console.log("Server ready on :3000"));
// GET /health → { "status": "ok", "timestamp": "..." }
// GET /ready  → { "status": "ready", "agents": ["bot"] }

Core Examples

Multi-Agent & Orchestration

Knowledge & RAG

Deployment & Transport

Monitoring & Cost

Advanced & Integrations

End-to-End Applications

Transport & Deployment Examples

Express REST Server

Streaming Responses

Socket.IO Gateway

Swagger UI

File Upload

Per-Request API Keys

Multiple Agents on One Server

Teams and Workflows via REST

CORS and Security

Health Check Endpoint

​Express REST Server

​Streaming Responses

​Socket.IO Gateway

​Swagger UI

​File Upload

​Per-Request API Keys

​Multiple Agents on One Server

​Teams and Workflows via REST

​CORS and Security

​Health Check Endpoint

Express REST Server

Streaming Responses

Socket.IO Gateway

Swagger UI

File Upload

Per-Request API Keys

Multiple Agents on One Server

Teams and Workflows via REST

CORS and Security

Health Check Endpoint