CodeSense – AI Transparency Tool

AI Transparency Tool

How Much of This Repository
Was Written by AI?

Paste any public GitHub repo and instantly see which files look AI-generated, which are human-written, and where AI helped the most.

🔒 Public repos only · 🚫 Code is never stored · ⚡ Free forever

Scan depth:

Try an example:

What you will discover

🔍

File-by-File Verdict

Every file scored on 10 forensic dimensions — see exactly which ones look AI-generated

🗺️

AI Contribution Map

Understand which areas of the codebase AI helped with — UI vs API vs core logic

🔎

Source Detection

Identifies patterns from ChatGPT, Copilot, Cursor, Next.js templates, and more

👤

Developer Profile

Analyze a GitHub user across multiple repos and generate a shareable AI coding profile

Sample output — AI Contribution Map

🎨 UI Components

72%

🔌 API / Routes

38%

🔧 Utilities

24%

⚡ Core Logic

11%

💡 AI was mainly used for UI scaffolding. Core business logic is human-written.

📁

—

Authorship distribution

AI score per file

What patterns did we find? (average across all files)

Each bar below shows how strongly that pattern leans towards AI or human writing. A higher score means the code looks more AI-generated in that area.

Red = Looks AI-generated (score above 65%)

Yellow = Could be either (score 35–65%)

Green = Looks human-written (score below 35%)

🔍 Detected Code Sources

We analyze each file for patterns that match known AI tools, framework templates, and community sources. This is based on coding style, structure, and API usage — not just generic AI detection.

Results for each file — click any file to see why