Copilot Runtime

The Copilot Runtime is the backend layer that connects your frontend application to your AI agents. It's set up during the quickstart and is the recommended way to use CopilotKit.

Setting Up the Runtime#

The runtime is a lightweight server endpoint that you add to your backend:

npm install @copilotkit/runtime

Here's a minimal example using Next.js. createCopilotRuntimeHandler returns a plain fetch handler, so the route is just two exports. It lives at a catch-all path and exports both verbs, so the runtime can serve its sub-routes (/info, agent runs, threads) rather than a single URL:

import {
  CopilotRuntime,
  createCopilotRuntimeHandler,
  InMemoryAgentRunner,
} from "@copilotkit/runtime/v2";

const runtime = new CopilotRuntime({
  agents: {
    // your agents go here
  },
  runner: new InMemoryAgentRunner(),
});

const handler = createCopilotRuntimeHandler({
  runtime,
  basePath: "/api/copilotkit",
});

export const GET = handler;
export const POST = handler;
export const PATCH = handler;
export const DELETE = handler;

With the route in place, GET /api/copilotkit/info returns a JSON description of the runtime and the agents it has registered. That route is how tooling — and the frontend's transport auto-detection — discovers your runtime, so it is the quickest way to confirm the endpoint is wired up.

Then point your frontend at the endpoint:

import { provideCopilotKit } from "@copilotkit/angular";

provideCopilotKit({
  runtimeUrl: "/api/copilotkit",
})

For setup with other backend frameworks (Express, NestJS, Node.js HTTP), see the quickstart.

Which name identifies an agent#

The name you use to address an agent from the frontend must equal a key of the runtime's agents map. That key is the only name the frontend can ask for. An agent's own name, id, or class name is never used for routing, and the two are free to differ.

const runtime = new CopilotRuntime({
  agents: {
    // `my_agent` is the key — the one string the frontend may ask for.
    my_agent: new HttpAgent({ url: "http://localhost:8000/" }),
  },
});

<copilot-chat agentId="my_agent" />

Most integrations write that key literally in the runtime route, as above, so the binding is visible in one file. Some derive it instead, and that is where the rule stops being obvious:

• Mastra — MastraAgent.getRemoteAgents and getLocalAgents both build the map from listAgents(), which is keyed by the record key in new Mastra({ agents: { ... } }), not by the agent's id. Given new Agent({ name: "My Agent" }) exported as myAgent and registered as agents: { myAgent }, the runtime key is myAgent.
• LangGraph — graphId is a separate binding, from the runtime to a key in your deployment's langgraph.json. It does not have to equal the runtime's agents-map key, and it is not the name the frontend asks for. The starter template happens to use sample_agent for both.

To read the registered keys directly, hit GET {runtimeUrl}/info. It returns the agents the runtime advertises, under exactly the names the frontend must use.

The Default Agent#

If you register an agent with the name "default", CopilotKit's prebuilt UI components will use it automatically without any additional configuration on the frontend. This is useful when you have one primary agent and don't want to specify an agentId everywhere.

const runtime = new CopilotRuntime({
  agents: {
    // Frontend APIs use this agent when no other agent id is selected.
    default: new HttpAgent({ url: "https://my-agent.example.com" }),
  },
});

When you register multiple agents, the "default" agent powers the chat unless a specific agent is selected. Other agents remain addressable through the frontend agent API.

What the Runtime Provides#

Authentication and Security#

The runtime runs on your server, which means agent communication stays server-side. This gives you a trusted environment to enforce authentication, validate requests, and keep API keys secure. When you use the runtime, safe defaults are put in place so your agent endpoints are not exposed to unauthenticated access.

AG-UI Middleware#

The AG-UI protocol supports a middleware layer (agent.use) for logging, guardrails, request transformation, and more. Because the runtime runs server-side, this middleware executes in a trusted environment where it cannot be tampered with by the client.

Agent Routing#

When you register multiple agents with the runtime, it handles discovery and routing automatically. Your frontend doesn't need to know the details of where each agent lives or how to reach it.

CopilotKit Intelligence#

Features like threads and the inspector are provided through the runtime and CopilotKit Intelligence. These give you conversation persistence and debugging capabilities out of the box.

Built-in Middleware#

The runtime supports two first-class middleware options you can enable directly on CopilotRuntime without calling .use() on each agent manually.

A2UI#

Pass a2ui: {} to automatically apply A2UIMiddleware to all registered agents:

const runtime = new CopilotRuntime({
  agents: { default: myAgent },
  a2ui: {}, // enables A2UI rendering for all agents
});

To scope it to specific agents only, pass an agents list:

a2ui: {
  agents: ["my-agent"];
}

On the frontend, the A2UI renderer activates automatically. Configure a2ui only when you want to override its defaults:

provideCopilotKit({
  runtimeUrl: "/api/copilotkit",
  a2ui: { theme: myCustomTheme },
})

mcpApps#

Pass mcpApps to configure MCP servers for all agents from a single place:

const runtime = new CopilotRuntime({
  agents: { default: myAgent },
  mcpApps: {
    servers: [
      { type: "http", url: "http://localhost:3108/mcp", serverId: "my-server" },
    ],
  },
});

Each server entry optionally accepts an agentId field to scope that server to a single agent. Without it, the server is available to all agents.

What If I Want to Connect to My AG-UI Agent Directly?#

CopilotKit is built on the AG-UI protocol, which is an open standard. If you want to connect your frontend directly to an AG-UI-compatible agent without the runtime, pass the agent instance in your frontend configuration:

import { HttpAgent } from "@ag-ui/client";
import { provideCopilotKit } from "@copilotkit/angular";

provideCopilotKit({
  selfManagedAgents: {
    "my-agent": new HttpAgent({
      url: "https://my-agent.example.com",
    }),
  },
})

There are important things to understand before going this route:

1. Authentication is your responsibility. When you use the Copilot Runtime, safe defaults are put in place so that your agent endpoints are not exposed to unauthenticated access. When you connect directly, it is entirely up to you to secure your agent endpoint and manage authentication.

2. Many ecosystem features won't work. The AG-UI protocol supports a middleware layer designed to run on the backend. Many features in the CopilotKit ecosystem depend on this server-side middleware. Without the runtime, these features — including threads and other capabilities — will not be available.

Comparison#