The graphql-ws package is a lightweight and fast WebSocket server and client for GraphQL subscriptions, designed to work with the GraphQL JavaScript reference implementation. It allows you to set up a full-duplex communication channel between the client and server to facilitate real-time data exchange using the GraphQL protocol.

What are graphql-ws's main functionalities?

Creating a WebSocket server for GraphQL subscriptions

This code sample demonstrates how to create a WebSocket server that handles GraphQL subscriptions. It uses the `useServer` function from `graphql-ws` to attach the WebSocket server to an existing HTTP server.

const { createServer } = require('http');
const { execute, subscribe } = require('graphql');
const { useServer } = require('graphql-ws/lib/use/ws');
const { schema } = require('./my-graphql-schema');

const server = createServer(function weServeSocketsOnly(_, res) {
  res.writeHead(404);
  res.end();
});

const wsServer = new WebSocket.Server({
  server,
  path: '/graphql',
});

useServer({ schema, execute, subscribe }, wsServer);

server.listen(4000);

Creating a WebSocket client for GraphQL subscriptions

This code sample shows how to create a WebSocket client that can subscribe to GraphQL subscriptions. The `createClient` function from `graphql-ws` is used to establish a connection to the WebSocket server, and the `subscribe` method is used to listen for real-time data updates.

const { createClient } = require('graphql-ws');

const client = createClient({
  url: 'ws://localhost:4000/graphql',
});

client.subscribe(
  {
    query: 'subscription { somethingChanged { id } }',
    variables: {},
  },
  {
    next: data => console.log('data received:', data),
    error: err => console.error('error during subscription:', err),
    complete: () => console.log('subscription completed'),
  }
);

Other packages similar to graphql-ws

GraphQLOverWebSocket

Coherent, zero-dependency, lazy, simple, GraphQL over WebSocket Protocol compliant server and client.

Getting started

Install

$ yarn add graphql-ws

Create a GraphQL schema

import { buildSchema } from 'graphql';

// Construct a schema, using GraphQL schema language
const schema = buildSchema(`
  type Query {
    hello: String
  }
  type Subscription {
    greetings: String
  }
`);

// The roots provide resolvers for each GraphQL operation
const roots = {
  query: {
    hello: () => 'Hello World!',
  },
  subscription: {
    greetings: async function* sayHiIn5Languages() {
      for (const hi of ['Hi', 'Bonjour', 'Hola', 'Ciao', 'Zdravo']) {
        yield { greetings: hi };
      }
    },
  },
};

Start the server

With ws

import ws from 'ws'; // yarn add ws
import { useServer } from 'graphql-ws/lib/use/ws';

const server = new ws.Server({
  port: 4000,
  path: '/graphql',
});

useServer(
  // from the previous step
  { schema, roots },
  server,
);

console.log('Listening to port 4000');

With uWebSockets.js

import uWS from 'uWebSockets.js'; // yarn add uWebSockets.js@uNetworking/uWebSockets.js#<tag>
import { makeBehavior } from 'graphql-ws/lib/use/uWebSockets';

uWS
  .App()
  .ws(
    '/graphql',
    makeBehavior(
      // from the previous step
      { schema, roots },
    ),
  )
  .listen(4000, (listenSocket) => {
    if (listenSocket) {
      console.log('Listening to port 4000');
    }
  });

With fastify-websocket

import Fastify from 'fastify'; // yarn add fastify
import fastifyWebsocket from 'fastify-websocket'; // yarn add fastify-websocket
import { makeHandler } from 'graphql-ws/lib/use/fastify-websocket';

const fastify = Fastify();
fastify.register(fastifyWebsocket);

fastify.get(
  '/graphql',
  { websocket: true },
  makeHandler(
    // from the previous step
    { schema, roots },
  ),
);

fastify.listen(4000, (err) => {
  if (err) {
    fastify.log.error(err);
    return process.exit(1);
  }
  console.log('Listening to port 4000');
});

Use the client

import { createClient } from 'graphql-ws';

const client = createClient({
  url: 'ws://welcomer.com:4000/graphql',
});

// query
(async () => {
  const result = await new Promise((resolve, reject) => {
    let result;
    client.subscribe(
      {
        query: '{ hello }',
      },
      {
        next: (data) => (result = data),
        error: reject,
        complete: () => resolve(result),
      },
    );
  });

  expect(result).toEqual({ hello: 'Hello World!' });
})();

// subscription
(async () => {
  const onNext = () => {
    /* handle incoming values */
  };

  let unsubscribe = () => {
    /* complete the subscription */
  };

  await new Promise((resolve, reject) => {
    unsubscribe = client.subscribe(
      {
        query: 'subscription { greetings }',
      },
      {
        next: onNext,
        error: reject,
        complete: resolve,
      },
    );
  });

  expect(onNext).toBeCalledTimes(5); // we say "Hi" in 5 languages
})();

Recipes

🔗 Client usage with Promise

import { createClient, SubscribePayload } from 'graphql-ws';

const client = createClient({
  url: 'ws://hey.there:4000/graphql',
});

async function execute<T>(payload: SubscribePayload) {
  return new Promise<T>((resolve, reject) => {
    let result: T;
    client.subscribe<T>(payload, {
      next: (data) => (result = data),
      error: reject,
      complete: () => resolve(result),
    });
  });
}

// use
(async () => {
  try {
    const result = await execute({
      query: '{ hello }',
    });
    // complete
    // next = result = { data: { hello: 'Hello World!' } }
  } catch (err) {
    // error
  }
})();

🔗 Client usage with AsyncIterator

import { createClient, SubscribePayload } from 'graphql-ws';

const client = createClient({
  url: 'ws://iterators.ftw:4000/graphql',
});

function subscribe<T>(payload: SubscribePayload): AsyncIterableIterator<T> {
  let deferred: {
    resolve: (done: boolean) => void;
    reject: (err: unknown) => void;
  } | null = null;
  const pending: T[] = [];
  let throwMe: unknown = null,
    done = false;
  const dispose = client.subscribe<T>(payload, {
    next: (data) => {
      pending.push(data);
      deferred?.resolve(false);
    },
    error: (err) => {
      throwMe = err;
      deferred?.reject(throwMe);
    },
    complete: () => {
      done = true;
      deferred?.resolve(true);
    },
  });
  return {
    [Symbol.asyncIterator]() {
      return this;
    },
    async next() {
      if (done) return { done: true, value: undefined };
      if (throwMe) throw throwMe;
      if (pending.length) return { value: pending.shift()! };
      return (await new Promise<boolean>(
        (resolve, reject) => (deferred = { resolve, reject }),
      ))
        ? { done: true, value: undefined }
        : { value: pending.shift()! };
    },
    async return() {
      dispose();
      return { done: true, value: undefined };
    },
  };
}

(async () => {
  const subscription = subscribe({
    query: 'subscription { greetings }',
  });
  // subscription.return() to dispose

  for await (const result of subscription) {
    // next = result = { data: { greetings: 5x } }
  }
  // complete
})();

🔗 Client usage with Observable

import { Observable } from 'relay-runtime';
// or
import { Observable } from '@apollo/client/core';
// or
import { Observable } from 'rxjs';
// or
import Observable from 'zen-observable';
// or any other lib which implements Observables as per the ECMAScript proposal: https://github.com/tc39/proposal-observable

const client = createClient({
  url: 'ws://graphql.loves:4000/observables',
});

function toObservable(operation) {
  return new Observable((observer) =>
    client.subscribe(operation, {
      next: (data) => observer.next(data),
      error: (err) => observer.error(err),
      complete: () => observer.complete(),
    }),
  );
}

const observable = toObservable({ query: `subscription { ping }` });

const subscription = observable.subscribe({
  next: (data) => {
    expect(data).toBe({ data: { ping: 'pong' } });
  },
});

// ⏱

subscription.unsubscribe();

🔗 Client usage with Relay

import { GraphQLError } from 'graphql';
import {
  Network,
  Observable,
  RequestParameters,
  Variables,
} from 'relay-runtime';
import { createClient } from 'graphql-ws';

const subscriptionsClient = createClient({
  url: 'ws://i.love:4000/graphql',
  connectionParams: () => {
    const session = getSession();
    if (!session) {
      return {};
    }
    return {
      Authorization: `Bearer ${session.token}`,
    };
  },
});

// yes, both fetch AND subscribe handled in one implementation
function fetchOrSubscribe(operation: RequestParameters, variables: Variables) {
  return Observable.create((sink) => {
    if (!operation.text) {
      return sink.error(new Error('Operation text cannot be empty'));
    }
    return subscriptionsClient.subscribe(
      {
        operationName: operation.name,
        query: operation.text,
        variables,
      },
      {
        ...sink,
        error: (err) => {
          if (err instanceof Error) {
            return sink.error(err);
          }

          if (err instanceof CloseEvent) {
            return sink.error(
              // reason will be available on clean closes
              new Error(
                `Socket closed with event ${err.code} ${err.reason || ''}`,
              ),
            );
          }

          return sink.error(
            new Error(
              (err as GraphQLError[]).map(({ message }) => message).join(', '),
            ),
          );
        },
      },
    );
  });
}

export const network = Network.create(fetchOrSubscribe, fetchOrSubscribe);

🔗 Client usage with urql

import { createClient, defaultExchanges, subscriptionExchange } from 'urql';
import { createClient as createWSClient } from 'graphql-ws';

const wsClient = createWSClient({
  url: 'ws://its.urql:4000/graphql',
});

const client = createClient({
  url: '/graphql',
  exchanges: [
    ...defaultExchanges,
    subscriptionExchange({
      forwardSubscription(operation) {
        return {
          subscribe: (sink) => {
            const dispose = wsClient.subscribe(operation, sink);
            return {
              unsubscribe: dispose,
            };
          },
        };
      },
    }),
  ],
});

🔗 Client usage with Apollo

import {
  ApolloLink,
  Operation,
  FetchResult,
  Observable,
} from '@apollo/client/core';
import { print, GraphQLError } from 'graphql';
import { createClient, ClientOptions, Client } from 'graphql-ws';

class WebSocketLink extends ApolloLink {
  private client: Client;

  constructor(options: ClientOptions) {
    super();
    this.client = createClient(options);
  }

  public request(operation: Operation): Observable<FetchResult> {
    return new Observable((sink) => {
      return this.client.subscribe<FetchResult>(
        { ...operation, query: print(operation.query) },
        {
          next: sink.next.bind(sink),
          complete: sink.complete.bind(sink),
          error: (err) => {
            if (err instanceof Error) {
              return sink.error(err);
            }

            if (err instanceof CloseEvent) {
              return sink.error(
                // reason will be available on clean closes
                new Error(
                  `Socket closed with event ${err.code} ${err.reason || ''}`,
                ),
              );
            }

            return sink.error(
              new Error(
                (err as GraphQLError[])
                  .map(({ message }) => message)
                  .join(', '),
              ),
            );
          },
        },
      );
    });
  }
}

const link = new WebSocketLink({
  url: 'ws://where.is:4000/graphql',
  connectionParams: () => {
    const session = getSession();
    if (!session) {
      return {};
    }
    return {
      Authorization: `Bearer ${session.token}`,
    };
  },
});

🔗 Client usage with custom retry timeout strategy

import { createClient } from 'graphql-ws';
import { waitForHealthy } from './my-servers';

const url = 'ws://i.want.retry:4000/control/graphql';

const client = createClient({
  url,
  retryWait: async function waitForServerHealthyBeforeRetry() {
    // if you have a server healthcheck, you can wait for it to become
    // healthy before retrying after an abrupt disconnect (most commonly a restart)
    await waitForHealthy(url);

    // after the server becomes ready, wait for a second + random 1-4s timeout
    // (avoid DDoSing yourself) and try connecting again
    await new Promise((resolve) =>
      setTimeout(resolve, 1000 + Math.random() * 3000),
    );
  },
});

🔗 Client usage with graceful restart

import { createClient, Client, ClientOptions } from 'graphql-ws';
import { giveMeAFreshToken } from './token-giver';

interface RestartableClient extends Client {
  restart(): void;
}

function createRestartableClient(options: ClientOptions): RestartableClient {
  let restartRequested = false;
  let restart = () => {
    restartRequested = true;
  };

  const client = createClient({
    ...options,
    on: {
      ...options.on,
      connected: (socket) => {
        options.on?.connected?.(socket);

        restart = () => {
          if (socket.readyState === WebSocket.OPEN) {
            // if the socket is still open for the restart, do the restart
            socket.close(4205, 'Client Restart');
          } else {
            // otherwise the socket might've closed, indicate that you want
            // a restart on the next connected event
            restartRequested = true;
          }
        };

        // just in case you were eager to restart
        if (restartRequested) {
          restartRequested = false;
          restart();
        }
      },
    },
  });

  return {
    ...client,
    restart: () => restart(),
  };
}

const client = createRestartableClient({
  url: 'ws://graceful.restart:4000/is/a/non-fatal/close-code',
  connectionParams: async () => {
    const token = await giveMeAFreshToken();
    return { token };
  },
});

// all subscriptions from `client.subscribe` will resubscribe after `client.restart`

🔗 Client usage with ping/pong timeout and latency metrics

import { createClient } from 'graphql-ws';

let activeSocket,
  timedOut,
  pingSentAt = 0,
  latency = 0;
createClient({
  url: 'ws://i.time.out:4000/and-measure/latency',
  keepAlive: 10_000, // ping server every 10 seconds
  on: {
    connected: (socket) => (activeSocket = socket),
    ping: (received) => {
      if (!received /* sent */) {
        pingSentAt = Date.now();
        timedOut = setTimeout(() => {
          if (activeSocket.readyState === WebSocket.OPEN)
            activeSocket.close(4408, 'Request Timeout');
        }, 5_000); // wait 5 seconds for the pong and then close the connection
      }
    },
    pong: (received) => {
      if (received) {
        latency = Date.now() - pingSentAt;
        clearTimeout(timedOut); // pong is received, clear connection close timeout
      }
    },
  },
});

🔗 Client usage with manual pings and pongs

import {
  createClient,
  Client,
  ClientOptions,
  stringifyMessage,
  PingMessage,
  PongMessage,
  MessageType,
} from 'graphql-ws';

interface PingerClient extends Client {
  ping(payload?: PingMessage['payload']): void;
  pong(payload?: PongMessage['payload']): void;
}

function createPingerClient(options: ClientOptions): PingerClient {
  let activeSocket: WebSocket;

  const client = createClient({
    disablePong: true,
    ...options,
    on: {
      connected: (socket) => {
        options.on?.connected?.(socket);
        activeSocket = socket;
      },
    },
  });

  return {
    ...client,
    ping: (payload) => {
      if (activeSocket.readyState === WebSocket.OPEN)
        activeSocket.send(
          stringifyMessage({
            type: MessageType.Ping,
            payload,
          }),
        );
    },
    pong: (payload) => {
      if (activeSocket.readyState === WebSocket.OPEN)
        activeSocket.send(
          stringifyMessage({
            type: MessageType.Pong,
            payload,
          }),
        );
    },
  };
}

🔗 Client usage in browser

<!DOCTYPE html>
<html>
  <head>
    <meta charset="utf-8" />
    <title>GraphQL over WebSocket</title>
    <script
      type="text/javascript"
      src="https://unpkg.com/graphql-ws/umd/graphql-ws.min.js"
    ></script>
  </head>
  <body>
    <script type="text/javascript">
      const client = graphqlWs.createClient({
        url: 'ws://umdfor.the:4000/win/graphql',
      });

      // consider other recipes for usage inspiration
    </script>
  </body>
</html>

🔗 Client usage in Node

const ws = require('ws'); // yarn add ws
const Crypto = require('crypto');
const { createClient } = require('graphql-ws');

const client = createClient({
  url: 'ws://no.browser:4000/graphql',
  webSocketImpl: ws,
  /**
   * Generates a v4 UUID to be used as the ID.
   * Reference: https://gist.github.com/jed/982883
   */
  generateID: () =>
    ([1e7] + -1e3 + -4e3 + -8e3 + -1e11).replace(/[018]/g, (c) =>
      (c ^ (Crypto.randomBytes(1)[0] & (15 >> (c / 4)))).toString(16),
    ),
});

// consider other recipes for usage inspiration

🔗 Server usage with ws

// minimal version of `import { useServer } from 'graphql-ws/lib/use/ws';`

import ws from 'ws'; // yarn add ws
import { makeServer } from 'graphql-ws';
import { schema } from './my-graphql-schema';

// make
const server = makeServer({ schema });

// create websocket server
const wsServer = new ws.Server({
  port: 4000,
  path: '/graphql',
});

// implement
wsServer.on('connection', (socket, request) => {
  // a new socket opened, let graphql-ws take over
  const closed = server.opened(
    {
      protocol: socket.protocol, // will be validated
      send: (data) =>
        new Promise((resolve, reject) => {
          socket.send(data, (err) => (err ? reject(err) : resolve()));
        }), // control your data flow by timing the promise resolve
      close: (code, reason) => socket.close(code, reason), // there are protocol standard closures
      onMessage: (cb) =>
        socket.on('message', async (event) => {
          try {
            // wait for the the operation to complete
            // - if init message, waits for connect
            // - if query/mutation, waits for result
            // - if subscription, waits for complete
            await cb(event.toString());
          } catch (err) {
            // all errors that could be thrown during the
            // execution of operations will be caught here
            socket.close(1011, err.message);
          }
        }),
    },
    // pass values to the `extra` field in the context
    { socket, request },
  );

  // notify server that the socket closed
  socket.once('close', (code, reason) => closed(code, reason));
});

🔗 Server usage with ws and custom auth handling

// check extended implementation at `{ useServer } from 'graphql-ws/lib/use/ws'`

import http from 'http';
import ws from 'ws'; // yarn add ws
import { makeServer } from 'graphql-ws';
import { schema } from './my-graphql-schema';
import { validate } from './my-auth';

// extra in the context
interface Extra {
  readonly request: http.IncomingMessage;
}

// your custom auth
class Forbidden extends Error {}
function handleAuth(request: http.IncomingMessage) {
  // do your auth on every subscription connect
  const good = validate(request.headers['authorization']);
  // or const { iDontApprove } = session(request.cookies);
  if (!good) {
    // throw a custom error to be handled
    throw new Forbidden(':(');
  }
}

// make graphql server
const gqlServer = makeServer<Extra>({
  schema,
  onConnect: async (ctx) => {
    // do your auth on every connect
    await handleAuth(ctx.extra.request);
  },
  onSubscribe: async (ctx) => {
    // or maybe on every subscribe
    await handleAuth(ctx.extra.request);
  },
  onNext: async (ctx) => {
    // haha why not on every result emission?
    await handleAuth(ctx.extra.request);
  },
});

// create websocket server
const wsServer = new ws.Server({
  port: 4000,
  path: '/graphql',
});

// implement
wsServer.on('connection', (socket, request) => {
  // you may even reject the connection without ever reaching the lib
  // return socket.close(4403, 'Forbidden');

  // pass the connection to graphql-ws
  const closed = gqlServer.opened(
    {
      protocol: socket.protocol, // will be validated
      send: (data) =>
        new Promise((resolve, reject) => {
          // control your data flow by timing the promise resolve
          socket.send(data, (err) => (err ? reject(err) : resolve()));
        }),
      close: (code, reason) => socket.close(code, reason), // for standard closures
      onMessage: (cb) => {
        socket.on('message', async (event) => {
          try {
            // wait for the the operation to complete
            // - if init message, waits for connect
            // - if query/mutation, waits for result
            // - if subscription, waits for complete
            await cb(event.toString());
          } catch (err) {
            // all errors that could be thrown during the
            // execution of operations will be caught here
            if (err instanceof Forbidden) {
              // your magic
            } else {
              socket.close(1011, err.message);
            }
          }
        });
      },
    },
    // pass request to the extra
    { request },
  );

  // notify server that the socket closed
  socket.once('close', (code, reason) => closed(code, reason));
});

🔗 ws server usage with Express GraphQL

import ws from 'ws'; // yarn add ws
import express from 'express';
import { graphqlHTTP } from 'express-graphql';
import { useServer } from 'graphql-ws/lib/use/ws';
import { schema } from './my-graphql-schema';

// create express and middleware
const app = express();
app.use('/graphql', graphqlHTTP({ schema }));

const server = app.listen(4000, () => {
  // create and use the websocket server
  const wsServer = new ws.Server({
    server,
    path: '/graphql',
  });

  useServer({ schema }, wsServer);
});

🔗 ws server usage with Apollo Server Express

import express from 'express';
import { ApolloServer } from 'apollo-server-express';
import ws from 'ws'; // yarn add ws
import { useServer } from 'graphql-ws/lib/use/ws';
import { schema } from './my-graphql-schema';

// create express
const app = express();

// create apollo server
const apolloServer = new ApolloServer({ schema });

// apply middleware
apolloServer.applyMiddleware({ app });

const server = app.listen(4000, () => {
  // create and use the websocket server
  const wsServer = new ws.Server({
    server,
    path: '/graphql',
  });

  useServer({ schema }, wsServer);
});

🔗 ws server usage with subscriptions-transport-ws backwards compatibility

import https from 'https';
import ws from 'ws'; // yarn add ws
import { execute, subscribe } from 'graphql';
import { GRAPHQL_TRANSPORT_WS_PROTOCOL } from 'graphql-ws';
import { useServer } from 'graphql-ws/lib/use/ws';
import { SubscriptionServer, GRAPHQL_WS } from 'subscriptions-transport-ws';
import { schema } from './my-graphql-schema';

// graphql-ws
const graphqlWs = new ws.Server({ noServer: true });
useServer({ schema }, graphqlWs);

// subscriptions-transport-ws
const subTransWs = new ws.Server({ noServer: true });
SubscriptionServer.create(
  {
    schema,
    execute,
    subscribe,
  },
  subTransWs,
);

// create https server
const server = https.createServer(function weServeSocketsOnly(_, res) {
  res.writeHead(404);
  res.end();
});

// listen for upgrades and delegate requests according to the WS subprotocol
server.on('upgrade', (req, socket, head) => {
  // extract websocket subprotocol from header
  const protocol = req.headers['sec-websocket-protocol'];
  const protocols = Array.isArray(protocol)
    ? protocol
    : protocol?.split(',').map((p) => p.trim());

  // decide which websocket server to use
  const wss =
    protocols?.includes(GRAPHQL_WS) && // subscriptions-transport-ws subprotocol
    !protocols.includes(GRAPHQL_TRANSPORT_WS_PROTOCOL) // graphql-ws subprotocol
      ? subTransWs
      : // graphql-ws will welcome its own subprotocol and
        // gracefully reject invalid ones. if the client supports
        // both transports, graphql-ws will prevail
        graphqlWs;
  wss.handleUpgrade(req, socket, head, (ws) => {
    wss.emit('connection', ws, req);
  });
});

🔗 ws server usage with console logging

import ws from 'ws'; // yarn add ws
import { useServer } from 'graphql-ws/lib/use/ws';
import { schema } from './my-graphql-schema';

const wsServer = new ws.Server({
  port: 4000,
  path: '/graphql',
});

useServer(
  {
    schema,
    onConnect: (ctx) => {
      console.log('Connect', ctx);
    },
    onSubscribe: (ctx, msg) => {
      console.log('Subscribe', { ctx, msg });
    },
    onNext: (ctx, msg, args, result) => {
      console.debug('Next', { ctx, msg, args, result });
    },
    onError: (ctx, msg, errors) => {
      console.error('Error', { ctx, msg, errors });
    },
    onComplete: (ctx, msg) => {
      console.log('Complete', { ctx, msg });
    },
  },
  wsServer,
);

🔗 ws server usage on a multi WebSocket server

import https from 'https';
import ws from 'ws'; // yarn add ws
import url from 'url';
import { createClient } from 'graphql-ws';
import { useServer } from 'graphql-ws/lib/use/ws';
import { schema } from './my-graphql-schema';

const server = https.createServer(function weServeSocketsOnly(_, res) {
  res.writeHead(404);
  res.end();
});

/**
 * Two websocket servers on different paths:
 * - `/wave` sends out waves
 * - `/graphql` serves graphql
 */
const waveWS = new ws.Server({ noServer: true });
const graphqlWS = new ws.Server({ noServer: true });

// delegate upgrade requests to relevant destinations
server.on('upgrade', (request, socket, head) => {
  const pathname = url.parse(request.url).pathname;

  if (pathname === '/wave') {
    return waveWS.handleUpgrade(request, socket, head, (client) => {
      waveWS.emit('connection', client, request);
    });
  }

  if (pathname === '/graphql') {
    return graphqlWS.handleUpgrade(request, socket, head, (client) => {
      graphqlWS.emit('connection', client, request);
    });
  }

  return socket.destroy();
});

// wave on connect
waveWS.on('connection', (socket) => {
  socket.send('🌊');
});

// serve graphql
useServer({ schema }, graphqlWS);

server.listen(4000);

🔗 ws server usage with custom context value

import ws from 'ws'; // yarn add ws
import { useServer } from 'graphql-ws/lib/use/ws';
import { schema, roots, getDynamicContext } from './my-graphql';

const wsServer = new ws.Server({
  port: 4000,
  path: '/graphql',
});

useServer(
  {
    context: (ctx, msg, args) => {
      return getDynamicContext(ctx, msg, args);
    }, // or static context by supplying the value direcly
    schema,
    roots,
  },
  wsServer,
);

🔗 ws server usage with dynamic schema

import ws from 'ws'; // yarn add ws
import { useServer } from 'graphql-ws/lib/use/ws';
import { schema, checkIsAdmin, getDebugSchema } from './my-graphql';

const wsServer = new ws.Server({
  port: 4000,
  path: '/graphql',
});

useServer(
  {
    schema: async (ctx, msg, executionArgsWithoutSchema) => {
      // will be called on every subscribe request
      // allowing you to dynamically supply the schema
      // using the depending on the provided arguments.
      // throwing an error here closes the socket with
      // the `Error` message in the close event reason
      const isAdmin = await checkIsAdmin(ctx.request);
      if (isAdmin) return getDebugSchema(ctx, msg, executionArgsWithoutSchema);
      return schema;
    },
  },
  wsServer,
);

🔗 ws server usage with custom validation

import ws from 'ws'; // yarn add ws
import { useServer } from 'graphql-ws/lib/use/ws';
import { validate } from 'graphql';
import { schema, myValidationRules } from './my-graphql';

const wsServer = new ws.Server({
  port: 4000,
  path: '/graphql',
});

useServer(
  {
    validate: (schema, document) =>
      validate(schema, document, myValidationRules),
  },
  wsServer,
);

🔗 ws server usage with custom execution arguments

import { parse, validate } from 'graphql';
import ws from 'ws'; // yarn add ws
import { useServer } from 'graphql-ws/lib/use/ws';
import { schema, myValidationRules } from './my-graphql';

const wsServer = new ws.Server({
  port: 4000,
  path: '/graphql',
});

useServer(
  {
    onSubscribe: (ctx, msg) => {
      const args = {
        schema,
        operationName: msg.payload.operationName,
        document: parse(msg.payload.query),
        variableValues: msg.payload.variables,
      };

      // dont forget to validate when returning custom execution args!
      const errors = validate(args.schema, args.document, myValidationRules);
      if (errors.length > 0) {
        return errors; // return `GraphQLError[]` to send `ErrorMessage` and stop subscription
      }

      return args;
    },
  },
  wsServer,
);

🔗 ws server usage accepting only subscription operations

import { parse, validate, getOperationAST, GraphQLError } from 'graphql';
import ws from 'ws'; // yarn add ws
import { useServer } from 'graphql-ws/lib/use/ws';
import { schema } from './my-graphql';

const wsServer = new ws.Server({
  port: 4000,
  path: '/graphql',
});

useServer(
  {
    onSubscribe: (_ctx, msg) => {
      // construct the execution arguments
      const args = {
        schema,
        operationName: msg.payload.operationName,
        document: parse(msg.payload.query),
        variableValues: msg.payload.variables,
      };

      const operationAST = getOperationAST(args.document, args.operationName);
      if (!operationAST) {
        // returning `GraphQLError[]` sends an `ErrorMessage` and stops the subscription
        return [new GraphQLError('Unable to identify operation')];
      }

      // handle mutation and query requests
      if (operationAST.operation !== 'subscription') {
        // returning `GraphQLError[]` sends an `ErrorMessage` and stops the subscription
        return [new GraphQLError('Only subscription operations are supported')];

        // or if you want to be strict and terminate the connection on illegal operations
        throw new Error('Only subscription operations are supported');
      }

      // dont forget to validate
      const errors = validate(args.schema, args.document);
      if (errors.length > 0) {
        // returning `GraphQLError[]` sends an `ErrorMessage` and stops the subscription
        return errors;
      }

      // ready execution arguments
      return args;
    },
  },
  wsServer,
);

🔗 ws server and client usage with persisted queries

// 🛸 server

import { parse } from 'graphql';
import ws from 'ws'; // yarn add ws
import { useServer } from 'graphql-ws/lib/use/ws';
import { schema } from './my-graphql-schema';

// a unique GraphQL execution ID used for representing
// a query in the persisted queries store. when subscribing
// you should use the `SubscriptionPayload.query` to transmit the id
type QueryID = string;

const queriesStore: Record<QueryID, ExecutionArgs> = {
  iWantTheGreetings: {
    schema, // you may even provide different schemas in the queries store
    document: parse('subscription Greetings { greetings }'),
  },
};

const wsServer = new ws.Server({
  port: 4000,
  path: '/graphql',
});

useServer(
  {
    onSubscribe: (_ctx, msg) => {
      const query = queriesStore[msg.payload.query];
      if (!query) {
        // for extra security you only allow the queries from the store
        throw new Error('404: Query Not Found');
      }
      return {
        ...query,
        variableValues: msg.payload.variables, // use the variables from the client
      };
    },
  },
  wsServer,
);

// 📺 client

import { createClient } from 'graphql-ws';

const client = createClient({
  url: 'ws://persisted.graphql:4000/queries',
});

(async () => {
  const onNext = () => {
    /**/
  };

  await new Promise((resolve, reject) => {
    client.subscribe(
      {
        query: 'iWantTheGreetings',
      },
      {
        next: onNext,
        error: reject,
        complete: resolve,
      },
    );
  });

  expect(onNext).toBeCalledTimes(5); // greetings in 5 languages
})();

🔗 ws server and client auth usage with token expiration, validation and refresh

// 🛸 server

import ws from 'ws'; // yarn add ws
import { useServer } from 'graphql-ws/lib/use/ws';
import { schema } from './my-graphql-schema';
import { isTokenValid } from './my-auth';

const wsServer = new WebSocket.Server({
  port: 4000,
  path: '/graphql',
});

useServer(
  {
    schema,
    onConnect: async (ctx) => {
      // do your auth check on every connect
      if (!(await isTokenValid(ctx.connectionParams?.token)))
        // returning false from the onConnect callback will close with `4403: Forbidden`;
        // therefore, being synonymous to ctx.extra.socket.close(4403, 'Forbidden');
        return false;
    },
    onSubscribe: async (ctx) => {
      // or maybe on every subscribe
      if (!(await isTokenValid(ctx.connectionParams?.token)))
        return ctx.extra.socket.close(4403, 'Forbidden');
    },
    onNext: async (ctx) => {
      // why not on every result emission? lol
      if (!(await isTokenValid(ctx.connectionParams?.token)))
        return ctx.extra.socket.close(4403, 'Forbidden');
    },
  },
  wsServer,
);

// 📺 client

import { createClient } from 'graphql-ws';
import {
  getCurrentToken,
  getCurrentTokenExpiresIn,
  refreshCurrentToken,
} from './my-auth';

// non-fatal WebSocket connection close events will cause the
// client to automatically reconnect. the retries are silent, meaning
// that the client will not error out unless the retry attempts have been
// exceeded or the close event was fatal (read more about the fatal
// close events in the documentation). additionally, all active subscriptions
// will automatically resubscribe upon successful reconnect. this behaviour
// can be leveraged to implement a secure and sound way of authentication;
// handling server-side validation, expiry indication and timely token refreshes

// indicates that the server closed the connection because of
// an auth problem. it indicates that the token should refresh
let shouldRefreshToken = false,
  // the socket close timeout due to token expiry
  tokenExpiryTimeout = null;

const client = createClient({
  url: 'ws://server-validates.auth:4000/graphql',
  connectionParams: async () => {
    if (shouldRefreshToken) {
      // refresh the token because it is no longer valid
      await refreshCurrentToken();
      // and reset the flag to avoid refreshing too many times
      shouldRefreshToken = false;
    }
    return { token: getCurrentToken() };
  },
  on: {
    connected: (socket) => {
      // clear timeout on every connect for debouncing the expiry
      clearTimeout(tokenExpiryTimeout);

      // set a token expiry timeout for closing the socket
      // with an `4403: Forbidden` close event indicating
      // that the token expired. the `closed` event listner below
      // will set the token refresh flag to true
      tokenExpiryTimeout = setTimeout(() => {
        if (socket.readyState === WebSocket.OPEN)
          socket.close(4403, 'Unauthorized');
      }, getCurrentTokenExpiresIn());
    },
    closed: (event) => {
      // if closed with the `4403: Forbidden` close event
      // the client or the server is communicating that the token
      // is no longer valid and should be therefore refreshed
      if (event.code === 4403) shouldRefreshToken = true;
    },
  },
});

Documentation

Check the docs folder out for TypeDoc generated documentation.

How does it work?

Read about the exact transport intricacies used by the library in the GraphQL over WebSocket Protocol document.

Want to help?

File a bug, contribute with code, or improve documentation? Read up on our guidelines for contributing and drive development with yarn test --watch away!

Disclaimer

This library and the GraphQL over WebSocket Protocol are not cross-compatible with the legacy subscriptions-transport-ws and its accompanying Protocol.

You must use graphql-ws coherently and implement the GraphQL over WebSocket Protocol on both sides, server and the client.

5.1.0 (2021-06-09)

Features

client: disablePong option for when implementing a custom pinger (6510360), closes #117
Optional payload for ping/pong message types (2fe0345), closes #117

Keywords

FAQs

What is graphql-ws?

Is graphql-ws popular?

Is graphql-ws well maintained?

Package last updated on 09 Jun 2021

Did you know?

Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.

Install

graphql-ws

What is graphql-ws?

What are graphql-ws's main functionalities?

Other packages similar to graphql-ws

subscriptions-transport-ws

apollo-server

Getting started

Install

Create a GraphQL schema

Start the server

With ws

With uWebSockets.js

With fastify-websocket

Use the client

Recipes

Documentation

How does it work?

Want to help?

Disclaimer

5.1.0 (2021-06-09)

Features

Keywords

Related posts

Node.js Implements Stricter Policies for Semver-Major Pull Requests Ahead of Release Deadlines

Roblox Developers Targeted with npm Packages Infected with Skuld Infostealer and Blank Grabber

vlt Debuts New JavaScript Package Manager and Serverless Registry at NodeConf EU