Today marks a major milestone for Vibe Code Detector. We are thrilled to announce the release of Version 2, a complete reimagining of how developers can analyze, understand, and improve the "vibe" of their code.
Why Vibe Check?
In the age of AI-assisted coding, the line between human-written and machine-generated code is blurring. While AI tools like GitHub Copilot and ChatGPT are incredible productivity boosters, they often leave behind subtle "fingerprints"—patterns of rigidity, over-explanation, and generic naming that can make a codebase feel soulless and harder to maintain.
We built Vibe Code Detector not to shame AI code, but to help developers maintain their unique voice. V2 takes this mission further with powerful new tools designed for real-world workflows.
What's New in V2?
1. GitHub Repository Analysis
You asked, we delivered. You can now analyze entire GitHub repositories directly. Simply paste a repo URL (e.g., github.com/facebook/react), and our engine will:
- Scan the file structure for organizational patterns.
- Sample key files to calculate an aggregate "Vibe Score".
- Identify "hotspots" of high AI probability.
Under the Hood
We use the GitHub API to fetch file trees and content. To respect rate limits and ensure speed, we perform a heuristic sampling of the most "impactful" files—typically core logic in src/ or lib/ folders—rather than scanning every single asset.
2. URL Analysis
Want to check a Gist, a Pastebin, or a documentation page? The new URL input type allows you to fetch and analyze code from any public URL. We handle the scraping and parsing so you just get the results.
3. The Comparison Tool
This is our favorite new feature. The /compare page allows you to put two snippets side-by-side. It's perfect for:
- A/B testing different prompts.
- Comparing your refactor against the original AI suggestion.
- Settling debates during code review.
The New Vibe Engine
V2 isn't just a UI facelift. We've overhauled our detection algorithm to look for deeper signals:
Identifier Entropy
We measure how "surprising" your variable names are. Humans tend to use domain-specific, creative names. AI defaults to data, item, and result.
Comment Density & Quality
It's not just about how many comments, but what they say. We look for "why" comments (human) vs. "what" comments (AI).
What's Next?
We're just getting started. In the coming weeks, look out for:
- VS Code Extension: Get vibe checks right in your editor.
- Team Dashboards: Track your team's vibe over time.
- API Access: Integrate vibe checking into your CI/CD pipeline.
Try out the new features today at vibecodedetector.com and let us know what you think!