I built SpectreAnalysis because of a problem I kept running into — I couldn't tell which parts of my code were written by a person and which were generated by AI. Not because the AI code was bad. It was fine. That was the thing. It blended right in. That got me thinking — if I can't tell in a single code file, what does that look like across an entire codebase? Across months? Across a whole team? I asked around. Everyone had a feeling. Nobody had a number. Meanwhile, engineering leaders are spending thousands on Copilot, Cursor, and other AI tools — and when the board asks "what's the ROI?" the honest answer is usually "we don't really know." So I built SpectreAnalysis. It connects to your GitHub repos, scans every PR, and gives you line-level attribution — what came from Copilot, Claude, Cursor, or a human. Here's what you get: - AI vs human code breakdown per repo, per developer, per PR - - Trend tracking over time - Policy enforcement (e.g. "no more than 60% AI code without review") - Team analytics and executive dashboards. It's free for up to 3 repositories. It's still early. I'm actively building and fixing things over weekends. If you find bugs, please drop them on GitHub — https://github.com/kloudd/Spectr... Would love to hear — how does your team measure AI tool ROI today?

Spectre

Know the ROI of your AI

Know the ROI of your AI