Performance
Transparent, reproducible benchmarks comparing Pyxle against popular web frameworks. Every framework implements identical endpoints with the same logic.
Tool
autocannon v8
Duration
12s per test
Connections
10 concurrent
Hardware
Apple M3 (8 cores)
All Python frameworks use uvicorn (single worker, ASGI) except Flask (gunicorn, 4 WSGI workers). Node.js frameworks use their default server. Pyxle runs in production mode via pyxle serve with full middleware stack. Database: SQLite with WAL mode, 1,000 pre-seeded rows. Source code (run it yourself)
Highlights
Pyxle outperforms Django by 2.2x on average across all API benchmarks, while offering SSR, file-based routing, and server actions that neither Django nor Flask provides.
Pyxle renders server-side pages at 1,100+ req/s with a persistent worker pool and bundle caching. The gap with Next.js is just 1.4x — remarkable for a Python + Node.js hybrid architecture.
Across all tests at 100 concurrent connections, Pyxle maintained zero errors and zero timeouts. Production-grade stability is not optional.
Pyxle handles POST/form processing at 20,000+ req/s — nearly double FastAPI. Starlette's ASGI layer combined with Pyxle's lean middleware stack shines for request body processing.
SSR Performance
Pyxle is the only Python framework that renders React on the server. Its optimized worker pool with bundle caching delivers throughput comparable to Next.js — the industry standard.
| Connections | Pyxle (req/s) | Next.js (req/s) | Ratio |
|---|---|---|---|
| 1 | 441 | 1,364 | 3.1x |
| 5 | 1,039 | 1,645 | 1.6x |
| 10 | 1,040 | 1,620 | 1.6x |
| 50 | 1,149 | 1,659 | 1.4x |
| 100 | 1,168 | 1,618 | 1.4x |
API Benchmarks
Each framework implements identical API endpoints. Pyxle runs its full middleware stack (SSR support, GZip compression, server actions). API-only frameworks run leaner stacks by design.
Parse JSON body, validate fields, return response. Measures request processing pipeline.
Return a static JSON object. Measures pure framework and serialization overhead.
Minimal endpoint. Measures raw framework routing overhead.
Read one random row from SQLite. Measures framework + database access overhead.
Read 5 random rows from SQLite. Measures query loop performance.
Read 20 random rows from SQLite. Heavier database workload.
Pyxle is a full-stack framework — every request passes through its SSR-ready middleware stack, GZip compression, and server action routing, even on API-only endpoints. Lightweight API frameworks like FastAPI and Flask skip that overhead entirely. For real-world applications, Pyxle's value is that you get SSR, server actions, and Python's ecosystem in a single framework — at competitive throughput.
All benchmark code is open source. Clone the repo, start the servers, and run the suite on your own hardware.
View benchmark source