🖥 Platform Support

Pre-compiled binaries for every platform and every language — no compiler required, works out of the box.

📦 Installation — 7 registries, 1 core

All seven packages ship the same native core (pre-compiled, no compiler needed). The Python wheel additionally bundles the HTTP middleware suite, and the Ruby/PHP/Go packages bundle the native lib for their platform.

Languages: العربية · Deutsch · Español · Français · हिन्दी · Italiano · 日本語 · 한국어 · Português (BR) · Русский · 中文

🧩 What each SDK gives you

The compression + minification + JSON core is the same native call in all seven languages (one C-ABI, contract-tested per language). The cache, bloom filter and circuit breaker are contract-tested in Python, Node.js, Java and .NET; their Ruby, PHP and Go wrappers are on the roadmap.

⚡ Performance at a Glance

🚀 Quick Start

Try it in one command (offline, no setup)

pip install b4n1-boost
b4n1-boost demo          # live numbers from the engine, on your machine
b4n1-boost docs --list   # the manual, bundled inside the binary

b4n1-boost demo runs the engine's own benchmark where it is executed and prints what it just measured — no stored figures, no projections, no network. The historical per-framework multipliers are printed as estimates, labelled as such.

Prefer a file you can hand over? bash scripts/build-demo-bundle.sh (or bash cli.sh boost bundle from the framework root) builds a self-contained ZIP with the pitch, the full manual, the binary and a one-command demo.sh.

The shared core (Python, Node.js, Java, .NET)

Response cache, bloom filter and circuit breaker — the same native primitives in these four SDKs. Contract-tested in all four.

Node.js

const { ResponseCache, BloomFilter, CircuitBreaker } = require('b4n1-boost/core.js');

const cache = new ResponseCache(1024);                       // capacity
cache.set('user:1', JSON.stringify({ id: 1 }), 200, 60);    // key, body, status, ttl (s)

const seen = new BloomFilter(100_000, 0.01);                // expected items, fp rate
if (!seen.checkAndSet('already-processed')) { /* first time we see it */ }

const breaker = new CircuitBreaker(5, 2, 30);               // threshold, successes, timeout (s)
if (!breaker.allows()) return fallback();

.NET

using B4N1.Boost;

using var cache = new NativeResponseCache(1024);
cache.Set("user:1", Encoding.UTF8.GetBytes("{\"id\":1}"), 200, 60);

using var seen = new NativeBloomFilter(100_000, 0.01);
bool firstTime = !seen.CheckAndSet("already-processed");

var breaker = new NativeCircuitBreaker(5, 2, 30);
if (!breaker.Allows) return Fallback();

Java / Kotlin

import com.b4n1.boost.NativeResponseCache;

try (NativeResponseCache cache = new NativeResponseCache(8)) {
    cache.set("user:1", "{\"id\":1}".getBytes(StandardCharsets.UTF_8), 200, 60_000);
}

Python

from b4n1_boost import ResponseCache

cache = ResponseCache(1024, 60)                   # capacity, default TTL (s)
cache.set("user:1", b'{"id":1}', [], 200)         # key, body, headers, status

Compression engine

Node.js

const boost = require('b4n1-boost');

const zstd = boost.compressZstd(payload);      // ~21x faster than gzip level 4 (1MB)
const back = boost.decompressZstd(zstd);
const min  = boost.minifyHtml(html);

.NET

var zstd = NativeBoost.CompressZstd(payload);

Java / Kotlin

byte[] zstd = Boost.compressZstd(data, 3);

Ruby

require 'boost'

zstd = B4N1::Boost.compress_zstd(payload)
min  = B4N1::Boost.minify_html(html)
ok   = B4N1::Boost.json_valid?(body)

PHP

use B4N1\Boost;

$zstd = Boost::compressZstd($payload);
$min  = Boost::minifyHtml($html);
$ok   = Boost::isJsonValid($body);

Go

import boost "github.com/B4N1-com/b4n1-boost/sdks/go"

zstd, _ := boost.CompressZstd(payload, 3)
min := boost.MinifyHTML(html)
ok := boost.IsJSONValid(body)

Express / ASP.NET middleware

const express = require('express');
const boostMiddleware = require('b4n1-boost/express');

app.use(boostMiddleware({ minifyHtml: true, gzipLevel: 4 }));

📚 Full usage for every language lives in the manual.

🐍 Python suite (middleware + accelerators)

The Python package is the most complete one: everything above, plus the HTTP middleware suite and the framework accelerators.

pip install b4n1-boost

Precompiled native wheels for Linux (x86_64 + aarch64), macOS (x86_64 + Apple Silicon), and Windows (x86_64). No compiler required.

Supported Python versions: 3.10, 3.11, 3.12, 3.13

Django

import b4n1_boost
report = b4n1_boost.install_django()
# Appends DjangoBoostMiddleware to settings.MIDDLEWARE
# Report shows: {'framework': 'Django', 'middleware_installed': True, ...}

FastAPI

from fastapi import FastAPI
import b4n1_boost

app = FastAPI()
b4n1_boost.install_fastapi(app)

Flask

from flask import Flask
import b4n1_boost

app = Flask(__name__)
b4n1_boost.install_flask(app)

Auto-detection

import b4n1_boost
b4n1_boost.autoboost()  # Detects Django/FastAPI/Flask automatically

One-line install all middleware

import b4n1_boost
report = b4n1_boost.boost_all()
# Applies: compression + ETag + security headers + CORS + rate limiting + health check
# Returns: {'framework': 'Django', 'applied': ['compression', 'etag', ...], 'middleware_count': 6}

🔧 Features

JSON Serialization (3.1-5.0x faster than stdlib json)

from b4n1_boost import NativeJson, canonicalize_json, validate_json

# Fast JSON dumps (native engine, 3.1-5.0x faster than stdlib `json`)
result = NativeJson.dumps({"users": [...]})

# Direct native path (no intermediate conversion)
result = NativeJson.dumps_direct(data)

# Batch: serialize N objects in a single GIL acquire
results = NativeJson.batch_dumps([obj1, obj2, obj3])

# Canonicalize: sorted keys, compact form (accepts str or dict)
canonicalize_json({"z": 1, "a": 2})  # '{"a":2,"z":1}'

# Fast JSON validation
validate_json('{"valid": true}')  # True

Compression (~21x faster than gzip level 4, GIL-free)

from b4n1_boost import py_compress_zstd, py_compress_gzip, py_compress_brotli

compressed = py_compress_zstd(payload)   # ~21x faster than gzip level 4 (1MB)
compressed = py_compress_brotli(payload) # Best ratio (slower than gzip)
compressed = py_compress_gzip(payload)   # ~1.0x vs stdlib gzip level 6

# Static file compression
from b4n1_boost import compress_static_file, compress_static_dir
compress_static_file("app/static/app.js", algorithm="zstd")
compress_static_dir("app/static/", algorithm="zstd")

Content-Type Aware Middleware

from b4n1_boost.middleware import B4N1BoostCompressionMiddleware

# Automatically skips: images, video, audio, fonts, archives, already-compressed
app.wsgi_app = B4N1BoostCompressionMiddleware(app.wsgi_app, min_size=1024)

# For FastAPI/ASGI (streaming support)
from b4n1_boost.middleware import FastAPIBoostCompressionMiddleware
app.add_middleware(FastAPIBoostCompressionMiddleware)

ETag / 304 Caching

from b4n1_boost.middleware import ETagMiddleware
app.wsgi_app = ETagMiddleware(app.wsgi_app)

Rate Limiting

from b4n1_boost.middleware import RateLimitMiddleware
app.wsgi_app = RateLimitMiddleware(app.wsgi_app, max_requests=100, window_seconds=60)

Security Headers

from b4n1_boost.middleware import SecurityHeadersMiddleware
app.wsgi_app = SecurityHeadersMiddleware(app.wsgi_app)
# Adds: X-Content-Type-Options, X-Frame-Options, HSTS, X-XSS-Protection, etc.

CORS

from b4n1_boost.middleware import CORSMiddleware
app.wsgi_app = CORSMiddleware(app.wsgi_app, allow_origins=["https://example.com"])

Health Check

from b4n1_boost.middleware import HealthCheckMiddleware
app.wsgi_app = HealthCheckMiddleware(app.wsgi_app, path="/health")
# GET /health → 200 {"status": "ok"}

Cache Layer

from b4n1_boost.advanced import ResponseCache

cache = ResponseCache(capacity=1000, default_ttl=300)
cache.set("/users", b'{"ok":true}', [], 200)
body, headers, status = cache.get("/users")

Django ORM Accelerator

from b4n1_boost.django_accelerator import bulk_insert_native, FastModelMixin

# PostgreSQL COPY — native bulk insert (speed not measured yet)
bulk_insert_native(MyModel, [
    {"name": "Alice", "email": "alice@example.com"},
    {"name": "Bob", "email": "bob@example.com"},
])

DRF Serializer Accelerator

from b4n1_boost.drf_accelerator import fast_serialize, FastSerializerMixin

# queryset → JSON without DRF overhead (speed not measured yet)
json_bytes = fast_serialize(queryset, fields=["id", "name", "email"])

# Mixin for existing serializers
class MySerializer(FastSerializerMixin, serializers.ModelSerializer):
    class Meta:
        model = MyModel
        fields = ["id", "name"]

JWT Validation

from b4n1_boost.advanced import validate_jwt

# Native HMAC-SHA256 (no PyJWT dependency required)
payload = validate_jwt(token, secret="my-secret", algorithm="HS256")

HTML/CSS/JS Minification

from b4n1_boost.advanced import minify_html, minify_css, minify_js

minified = minify_html("<html>  <body>  Hello  </body>  </html>")

HTML → Markdown (Agentic)

from b4n1_boost import html_to_markdown

markdown = html_to_markdown("<h1>Title</h1><p>Content with <b>bold</b></p>")
# → "# Title\n\nContent with **bold**"

Telemetry

from b4n1_boost import (
    telemetry_init,
    telemetry_capture_error,
    telemetry_log_info,
    telemetry_log_phase,
    telemetry_flush,
)

telemetry_init(app="b4n1-boost")
telemetry_capture_error("something failed", {"user": "123"})
telemetry_log_info("Request processed", {"phase": "http"})
telemetry_log_phase("http", "ok", 12)
telemetry_flush()

Background Worker (zero-GIL)

from b4n1_boost import (
    worker_compress,
    worker_decompress,
    worker_minify,
    worker_jwt_validate,
)

# Heavy ops offloaded to the background worker — the caller does not hold
# the GIL while they run
compressed = worker_compress(data, "zstd")
restored = worker_decompress(compressed)
markup = worker_minify(html, "html")
claims = worker_jwt_validate(token, secret)

🔍 Status & Diagnostics

import b4n1_boost
print(b4n1_boost.status())
# {'version': '0.4.3', 'native_extension': True, 'features': [...]}

# Run hardware benchmarks
report = b4n1_boost.run_benchmarks(iterations=100_000)

# Generate fresh HTML status report
# python3 generate_report.py --open

🔗 Links

• Website: https://b4n1.com
• PyPI: https://pypi.org/project/b4n1-boost
• npm: https://www.npmjs.com/package/b4n1-boost
• NuGet: https://www.nuget.org/packages/B4N1.Boost
• Maven Central: https://central.sonatype.com/artifact/com.b4n1/boost
• RubyGems: https://rubygems.org/gems/b4n1-boost
• Packagist: https://packagist.org/packages/b4n1/boost
• Go: https://pkg.go.dev/github.com/B4N1-com/b4n1-boost/sdks/go
• Repository: https://github.com/B4N1-com/b4n1-boost
• Licensing: https://b4n1.com/licensing or b4n1@b4n1.com
• Changelog: CHANGELOG.md

📄 License

Business Source License 1.1 (BUSL-1.1) — note this is not the SPDX id BSL-1.1, which denotes the Boost Software License.

• Free for development, evaluation, testing, personal projects, and startups under $100K USD annual revenue.
• Commercial license required for organizations >= $100K USD, government agencies, and public bidding.
• After Change Date (4 years) → Apache License 2.0.

See LICENSE for full text.

b4n1-boost: The polyglot accelerator. One native core, seven SDKs. JSON 3-5x faster. Compression ~21x faster than gzip level 4. Middleware transparent. Built with ❤️ by B4N1.