Skip to content
| Marketplace
Sign in
Visual Studio Code>Linters>Code Crawler - AI Code DetectionNew to Visual Studio Code? Get it now.
Code Crawler - AI Code Detection

Code Crawler - AI Code Detection

Not-for-vibe-coders

| (0) | Free
Detect AI-written vs human-written code with real-time keystroke tracking
Installation
Launch VS Code Quick Open (Ctrl+P), paste the following command, and press enter.
Copied to clipboard
More Info

Code Crawler - AI Code Detection

Detect AI-written vs human-written code in real time, right in your editor.

Code Crawler tracks your keystrokes and editing patterns to distinguish code you wrote yourself from code generated by AI tools like GitHub Copilot, Cursor, ChatGPT, and others.

Features

🔍 Real-Time Detection

Code Crawler monitors 11 behavioral signals as you type — typing rhythm, cursor movement, edit velocity, undo patterns, paste detection, and more — to determine whether each line was written by a human or generated by AI.

🎨 Line-Level Decorations

Every tracked line gets a color-coded gutter indicator:

  • 🟥 Red — AI-generated code
  • 🟩 Green — Human-written code
  • 🟨 Yellow — Mixed authorship

Confidence levels are shown via opacity: high (>80%), medium (50-80%), low (<50%).

💬 Hover Tooltips

Hover over any line to see the authorship verdict, confidence percentage, and which detection signal was most influential.

📊 Status Bar

A persistent status bar item shows the current file's AI vs Human percentages at a glance. Click it to open the full report.

📈 Sidebar Panel

An activity bar panel displays a pie chart breakdown of the current file's authorship, with line counts and confidence scores.

📋 Full Report

Run Code Crawler: Show Report from the Command Palette to generate a detailed markdown report of all tracked files in your workspace, including per-file signal breakdowns and reliability scores.

How It Works

Code Crawler uses a multi-signal fusion engine with 11 detection heuristics:

Signal What It Detects
Typing Rhythm Burst patterns, inter-keystroke timing
Edit Size Large insertions vs character-by-character typing
Cursor Movement Jump patterns vs sequential movement
Edit Sequence File navigation patterns
Undo Frequency Backtracking (highly human)
Extension Source AI extension command detection
Velocity Profile Typing speed variation
Selection Pattern Select-and-replace patterns
Paste Pattern Clipboard paste detection
Deletion Pattern Code removal patterns
Pause Pattern Thinking pauses between edits

Signals are weighted and fused using a confidence engine. Multiple corroborating signals increase certainty; contradictory signals reduce it.

Settings

Setting Default Description
codeCrawler.enabled true Enable or disable tracking
codeCrawler.aiThreshold 0.5 AI detection threshold (0–1)
codeCrawler.showDecorations true Show line-level authorship colors

Commands

Command Description
Code Crawler: Show Report Generate a full markdown report for all tracked files
Code Crawler: Reset Tracking Clear all authorship tracking data

Supported Languages

TypeScript, JavaScript, Python, Ruby, Go, Rust, Java, C, C++, C#, PHP, Swift — and any other language supported by VS Code.

Data & Privacy

All tracking data is stored locally in a .code-crawler/ directory within your workspace. No data is sent to any external servers.

License

MIT

  • Contact us
  • Jobs
  • Privacy
  • Manage cookies
  • Terms of use
  • Trademarks
© 2026 Microsoft