High-performance, TypeScript-first HTTP/HTTPS server built directly on raw TCP sockets for maximum throughput, and zero runtime dependencies. Supports HTTP/2 with automatic HTTP/1.1 fallback, pooled contexts for minimal GC, and a robust middleware pipeline.
Designed to integrate seamlessly into the ContextJS ecosystem or run standalone.
- Installation
- Features
- Benchmarks
- Quick Start
- Basic Middleware Example
- Streaming a File
- Configuration Reference
- API Reference
- Events
- Exceptions
npm i @contextjs/webserver
- Socket-based core leveraging raw TCP sockets for lowest-level performance
- HTTP and HTTPS support with PEM-encoded certificates
- Middleware pipeline for modular request handling
-
Context pooling for reusing
HttpContext
instances - Configurable header limits and idle socket cleanup
-
Lifecycle events (
info
,warning
,error
) for observability - TypeScript declarations with full JSDoc support
- Zero dependencies for maximum performance and minimal footprint
Below are the results of our WebServer benchmark suite, which compares throughput and latency across four different Node.js–based HTTP servers. Each test is run on GitHub Actions' provided "ubuntu-latest" runner. The benchmarks target a minimal server that responds with a 200 OK status and a short body of "OK" They use 500 concurrent connections, a pipelining factor of 1, and run for 10 seconds, with a warmup phase, and results averaged over three runs.
Server | Req/sec | Latency (ms) | Throughput (MB/s) | Errors |
---|---|---|---|---|
ContextJS | 15492.53 | 31.97 | 2.76 | 0.00 |
Node.js Raw HTTP | 14793.33 | 33.51 | 2.64 | 0.00 |
Fastify | 13884.27 | 35.80 | 2.48 | 0.00 |
Express | 8942.94 | 55.93 | 2.14 | 0.00 |
Column descriptions:
- Req/sec — Average number of HTTP requests served per second.
- Latency (ms) — Median (50th percentile) response time in milliseconds.
- Throughput (MB/s) — Average data transferred per second.
- Errors — Total connection-level failures (e.g. resets, refusals).
Server | Connections | Pipelining | Duration (s) | Latency Stdev (ms) | Requests Stdev | Throughput Stdev (MB/s) | Total Requests |
---|---|---|---|---|---|---|---|
ContextJS | 500 | 1 | 10.07 | 102.32 | 258.50 | 0.05 | 464750 |
Node.js Raw HTTP | 500 | 1 | 10.08 | 109.54 | 198.34 | 0.04 | 443750 |
Fastify | 500 | 1 | 10.10 | 121.67 | 174.27 | 0.03 | 416500 |
Express | 500 | 1 | 10.15 | 237.23 | 191.14 | 0.05 | 268250 |
Extended column descriptions:
- Connections — Number of simultaneous TCP connections opened.
- Pipelining — Number of requests pipelined per connection.
- Duration (s) — Total benchmark runtime in seconds.
- Latency Stdev (ms) — Standard deviation of response latency.
- Requests Stdev — Standard deviation of the requests/sec across samples.
- Throughput Stdev (MB/s) — Standard deviation of throughput.
- Total Requests — Sum of all successful 2xx responses across all iterations.
import { HttpContext, HttpWebServerOptions, WebServer, WebServerOptions } from "@contextjs/webserver";
const options = new WebServerOptions(
undefined,
new HttpWebServerOptions(true, "0.0.0.0", 3000, 60_000)
);
const server = new WebServer(options);
// Simple middleware
server.useMiddleware({
name: "hello-world",
async onRequest(context: HttpContext) {
await context.response
.setHeader("Content-Type", "text/plain")
.sendAsync("Hello, ContextJS!");
}
});
server.startAsync()
setTimeout(async () => {
await server.stopAsync();
console.log("Server stopped after 60 seconds");
}, 60000);
If you’re using the ContextJS Application
from @contextjs/system
, you can wire up the server directly:
import "@contextjs/webserver"; // module augmentation
import { Application } from "@contextjs/system";
import { HttpContext } from "@contextjs/webserver";
const app = new Application();
app.useWebServer(options => {
options.http.port = 8080;
options.onEvent = e => console.log(`[WebServer:${e.type}]`, e.detail);
});
app.webServer.useMiddleware({
name: "logger",
version: "1.0.0",
onRequest(context: HttpContext) {
console.log(`${context.request.method} ${context.request.path}`);
context.response
.setHeader("Content-Type", "text/plain")
.setHeader("X-ContextJS", "Hello World")
.sendAsync("Hello, ContextJS!");
}
});
await app.runAsync();
app.webServer.useMiddleware({
name: "static-file",
version: "1.0.0",
async onRequest(context: HttpContext, next: () => Promise<void>) {
if (context.request.path.startsWith("/assets/")) {
const filePath = path.join(__dirname, context.request.path);
return await context.response
.setHeader("Content-Type", "application/octet-stream")
.streamAsync(createReadStream(filePath));
}
await next();
}
});
import { HttpContext } from "@contextjs/webserver";
import { Controller, IActionResult, Ok, Post } from "@contextjs/webserver-middleware-controllers";
import fs from "node:fs";
import path from "node:path";
import { pipeline } from "node:stream/promises";
@Controller()
export class HomeController {
@Post("index")
public async indexPost(context: HttpContext): Promise<IActionResult> {
// Choose upload directory and file name
const uploadDir = path.join(process.cwd(), "uploads");
const fileName = "uploaded-file.png";
const uploadPath = path.join(uploadDir, fileName);
// Ensure upload directory exists
await fs.promises.mkdir(uploadDir, { recursive: true });
// Stream the request body directly to disk (no memory buffering)
const writeStream = fs.createWriteStream(uploadPath);
await pipeline(context.request.body, writeStream);
return Ok("File uploaded successfully!");
}
}
Property | Type | Description |
---|---|---|
general |
GeneralWebServerOptions |
Header limits, pool size, idle socket timeout |
http |
HttpWebServerOptions |
HTTP binding: port, host, keep-alive |
https |
HttpsWebServerOptions |
HTTPS binding: port, host, keep-alive, SSL certificate |
onEvent |
(event: WebServerEvent) |
Callback for info / warning / error events |
-
maximumHeaderSize: number
— max header bytes (default: 32 * 1024) -
httpContextPoolSize: number
— pre-allocate contexts (default: 1024) -
idleSocketsTimeout: number
— ms before closing idle sockets (default: 5000)
-
enabled: boolean
— enable HTTP (default: true) -
host?: string
— bind address (default: "localhost") -
port: number
— port number (default: 80) -
keepAliveTimeout: number
— ms for connection keep-alive (default: 5000)
-
enabled: boolean
— enable HTTPS (default: false) -
host?: string
— bind address (default: "localhost") -
port: number
— port number (default: 443) -
certificate: { key: string; cert: string }
— PEM key & cert -
keepAliveTimeout: number
— ms for connection keep-alive (default: 5000)
For detailed API documentation, please refer to the API Reference.
The server emits runtime events via the onEvent
callback:
-
info
— general progress messages -
warning
— recoverable issues (e.g. idle socket timeout) -
error
— fatal or unexpected errors