Performance comparison of Vercube against other popular Node.js frameworks.
When choosing a framework, performance is crucial. These benchmarks help you understand:
- How Vercube performs compared to established solutions
- Real-world metrics that impact your application (not synthetic tests)
- Trade-offs between different approaches (decorator-based, DI containers, etc.)
All frameworks are tested with identical endpoints and configuration for fair comparison.
- 🚀 Vercube (vercube.dev) - Modern, fast, TypeScript-first framework
- NestJS (nestjs.com) - Progressive Node.js framework
- Routing Controllers (github.com/typestack/routing-controllers) - Decorator-based routing
- Ts.ED (https://tsed.dev) - A modern framework written in TypeScript
- Rikta (https://rikta.dev/) - The Zero-Config TypeScript Framework
Last updated: Friday, January 16, 2026 at 2:17 PM UTC
Statistical comparison powered by hyperfine
| Framework | Mean | Median | Min | Max | vs Best |
|---|---|---|---|---|---|
| vercube 🏆 | 0.28s | 0.28s | 0.27s | 0.30s | — |
| routing-controllers | 0.42s | 0.42s | 0.41s | 0.43s | +49% |
| tsed | 0.46s | 0.46s | 0.44s | 0.49s | +62% |
| rikta | 0.97s | 0.97s | 0.93s | 1.04s | +244% |
| nestjs | 1.30s | 1.30s | 1.27s | 1.31s | +358% |
| Framework | Mean | Median | Min | Max | vs Best |
|---|---|---|---|---|---|
| vercube 🏆 | 280ms | 279ms | 255ms | 321ms | — |
| rikta | 326ms | 326ms | 306ms | 350ms | +16% |
| routing-controllers | 329ms | 331ms | 299ms | 381ms | +18% |
| nestjs | 377ms | 376ms | 350ms | 410ms | +35% |
| tsed | 946ms | 952ms | 899ms | 1021ms | +238% |
| Framework | Requests/sec | Latency p50 | Latency p95 | Latency p99 | vs Best RPS | vs Best p95 |
|---|---|---|---|---|---|---|
| vercube 🏆 | 95588 | 7.00ms | 19.00ms | 20.00ms | — | — |
| nestjs | 82705 | 9.00ms | 19.00ms | 20.00ms | -16% | +0% |
| rikta | 81156 | 9.00ms | 23.00ms | 25.00ms | -18% | +21% |
| routing-controllers | 78195 | 10.00ms | 20.00ms | 21.00ms | -22% | +5% |
| tsed | 32156 | 27.00ms | 56.00ms | 62.00ms | -197% | +195% |
Test environment: macOS 26.2, Apple M4 Pro (12 cores), 48GB RAM, Node.js 22
Load test config: 100 concurrent connections, 30s duration, 10 pipelining
- Bun - For running benchmark scripts (bun.sh)
- Node.js 22+ - For running framework applications
- pnpm 10+ - Package manager
# Clone and setup
git clone https://github.com/vercube/benchmarks
cd benchmark
pnpm run benchmark:setup
# Run complete benchmark suite
pnpm run benchmark
# View results
cat results/summary.md| Metric | Tool | Why It Matters |
|---|---|---|
| Build Time | hyperfine | Faster builds = better DX and CI/CD |
| Cold Start | hyperfine + custom script | Critical for serverless/edge deployments |
| Requests/sec | autocannon | Higher RPS = better scalability |
| Latency p95 | autocannon | 95% of users experience this or better |
| Memory/CPU | Custom Node.js monitor | Lower usage = lower hosting costs |
| Metric | Excellent | Good | Acceptable |
|---|---|---|---|
| Build | < 5s | 5-15s | > 15s |
| Cold Start | < 100ms | 100-500ms | > 500ms |
| RPS | > 50k | 20-50k | < 20k |
| Latency p95 | < 10ms | 10-50ms | 50-100ms |
| Memory | < 50MB | 50-100MB | 100-200MB |
Results vary based on hardware. See test environment details above.
For consistent, isolated results:
docker-compose up benchmarkBenefits: Same Node.js version, enforced resource limits (2 CPUs, 2GB RAM), reproducible across machines.
Want to add Hono, Fastify, Elysia, or improve the benchmarks? See CONTRIBUTING.md for detailed instructions.
- All solutions tested on identical hardware
- All use simple JSON endpoints to isolate framework overhead
- All use recommended production configurations
- Results reflect framework performance, not application code
For production decisions: Consider your specific use case, team experience, and ecosystem alongside these metrics.
Vercube is a modern, TypeScript-first Node.js framework designed for performance and developer experience.
MIT
Benchmarking tools: hyperfine, autocannon • Frameworks: Vercube, NestJS, Routing Controllers, Ts.ED, Rikta