Architectural Whitepaper & Technical Specifications

Engineering a Zero-Runtime Culinary Platform for the 2026/2027 Web

How Mise achieves sub-millisecond client responsiveness, zero runtime server dependencies, deterministic fractional mathematics, and resilient offline utility for professional home kitchens.

PageSpeed: 100/100
CLS: <0.005
TBT: 0ms
Agentic Discovery: 4/4 Pass
Runtime: Static Edge (Cloudflare)
View Portfolio Embed CardExplore Live CatalogHardware Diagnostics HUD
Interactive Data Flow

System Architecture & Data Pipeline

Select any pipeline stage below to inspect technical latency, protocols, security boundaries, and live implementation source code.

Static Export (SSG)

Next.js 16 Turbopack (output: "export")

All 12 recipes, catalog indexes, metadata routes, and case study pages are compiled into pure static HTML, CSS, and optimized JavaScript chunks during build time. Zero runtime Node.js server dependencies eliminates server cold-starts and vulnerable backend attack surfaces.

HTML Output

16 Prerendered Routes

Server Memory

0 MB Runtime

Cold Start

0.00 ms

Production Implementation CodeTypeScript Strict Mode
// next.config.ts
const nextConfig: NextConfig = {
  output: 'export', // Cloudflare Pages static export
  trailingSlash: true,
  images: { unoptimized: true },
};
export default nextConfig;
Under The Hood

Deep-Dive Technical Engineering Cards

Detailed breakdowns of four fundamental computer science and web engineering challenges solved in the Mise codebase.

Precision Culinary Mathematics

IEEE 754 Resolution

Standard JavaScript numbers use IEEE 754 64-bit floating point, leading to ugly representation issues (e.g. 1/3 serving yielding 0.3333333333333333 cups). Mise implements an exact Fractional Denomination Bucketing algorithm ($1/8, 1/4, 1/3, 1/2, 2/3, 3/4, 7/8$) with an epsilon threshold of $\pm 0.025$, ensuring recipes display clean kitchen fractions.

Float: 0.333

Culinary Fraction

1/3 cup

Metric Conversion

80 ml

Unit: fraction.ts15 Unit Tests Passing

Browser Hardware Integration

W3C Web APIs

Cooking at a kitchen counter with wet hands means cooks cannot unlock screens. Mise pairs the Screen Wake Lock API with automatic document visibilitychange re-acquisition to prevent phone displays from sleeping. Audio alerts avoid MP3/WAV network requests entirely by synthesizing a warm harmonic D5/A5 chime directly through the Web Audio API.

Web Audio Tone Synthesizer:587.33Hz (D5) → 880Hz (A5)

Zero Asset Chime Alert

Harmonic sine oscillator with exponential decay.

APIs: wakeLock & AudioContext0 Byte Audio Assets

Relational Set Matching

Fuzzy Culinary Lemmatization

Natural kitchen pantries rarely match exact recipe titles (e.g. "Eggs" vs "Free-Range Large Eggs (separated)"). Mise features a 4-tier culinary lemma normalizer: stripping adjectives (fresh, coarse, mature), removing temperature clauses (at 28°C), cleaning parenthesized specs, and stemming plurals into singular lemmas for token set intersection.

Toggle Pantry Items & Watch Match Score:

Sourdough Boule Match:
75%
Engine: pantryMatcher.ts14 Test Cases Passing

Autonomous Agentic Web Discovery

LLM Readability

Modern AI agents struggle with client-side Single Page Apps where data is locked inside DOM hydration. Mise exposes standardized /llms.txt and /llms-full.txt manifestos adhering to the agentic web spec, while embedding complete schema.org/Recipe JSON-LD graphs into every static HTML document.

/llms.txt

Route index & mathematical capability summary.

/llms-full.txt

Full Markdown corpus of all 12 platform recipes.

View /llms.txt fileJSON-LD 100% Valid
Standard: schema.org/RecipeDeterministic Agent Scrapes
Production Verification

Core Web Vitals & Cloudflare Edge Audit

Explore Recipe Catalog

100

Performance

0.6s LCP / 0ms TBT

100

Accessibility

WAI-ARIA 1.2 Full Pass

100

Best Practices

Zero Console Warnings

100

SEO

Semantic JSON-LD

Automated Deployment Pipeline (package.json)

$npm run test
$npm run typecheck
$npm run build:export
$npx wrangler pages deploy out