🔬 Public Benchmarks

Public Codebase Audit Benchmarks

Explore how famous open-source frameworks and AI starter templates score on our AST static engine.

★ Featured Benchmark Reports

Updated Weekly
Verified Open-Source
Soul:89%

React 19 Source Code

Grade A

14,850 lines of code across 142 files | 0 Hallucinations

Unmistakably Human. Highly contextual naming, zero redundant boilerplate.

Contextual Naming
94%
Boilerplate Entropy
Low
Creativity Index
High
// From React 19 Fiber Scheduler
let didWarnAboutMaps = false;
let didWarnAboutGenerators = false;

// Human pattern: conversational state flags with pragmatic domain context.
View Audit Breakdown→
AI-Generated Template
Soul:34%

v0 / Tailwind SaaS Scaffold

Grade C

3,420 lines of code across 28 files | 2 Unused Dependencies

AI-Generated. 12 code smells flagged, generic variable entropy, duplicate try-catch blocks.

Contextual Naming
41%
Boilerplate Entropy
High
Code Smells
12 Flagged
// From generated auth handler
try {
  const result = await processAuth(data);
  return result;
} catch (error: any) {
  console.log("Error in auth:", error);
  throw error;
} // Redundant pass-through try/catch boilerplate
View Audit Breakdown→
Human Core + AI Assist
Soul:82%

Supabase JS SDK

Grade A-

8,920 lines of code across 64 files | 0 Hallucinations

Human Core with AI Assist. Clean domain abstractions, strong TypeScript contracts.

Type Strictness
98%
Domain Abstraction
High
Documentation Ratio
Optimal
// From Supabase Client Builder
export class SupabaseClient<Database = any, SchemaName extends string & keyof Database = 'public'> {
  protected rest: PostgrestClient<Database, SchemaName>
  // Clean generics and explicit module boundary abstractions.
View Audit Breakdown→
🚀 App Launch Readiness

How Does Your App Compare Before Launch?

Run a 10-second AST audit on your public or private GitHub repository to verify security, performance, and code quality before going live.