小红书

Role

Product Designer & Developer

Timeline

Open-source web app · 2026

Outcome

Turns public star history into an evidence-backed investigation—without claiming causation.

Tools
Next.js
TypeScript
GitHubGitHub
GitHighlights — Read the code behind the curve
View on GitHub Try Live Demo

I used affaan-m/ECC as the investigation example.

10-SECOND OVERVIEW

From Star Spike to Git Evidence

Paste any public GitHub repository. GitHighlights finds unusual growth episodes and shows what changed in the code before each spike.

01

TRACE

Reconstruct the repository’s daily star history

02

DETECT

Find acceleration against its own historical baseline

03

INVESTIGATE

Review nearby commits, releases, files, and code churn

THE PRINCIPLEEvidence near the curve—not a causal claim.
PRODUCT TOUR

Follow the Curve Back to the Code

01 · START
Enter any public repository

Paste a GitHub URL or owner/repository pair. Example repositories make the first run immediate.

02 · FIND
Spot the moments growth accelerated

Detected episodes are layered over the cumulative star history and ranked for investigation.

03 · INVESTIGATE
Read the Git story around the spike

Inspect commits, releases, changed files, code churn, confidence, and the original evidence.

HOW IT WORKS

Useful Because It Knows Its Limits

01

OWN BASELINE

Each repository is compared with its own past—not a universal popularity threshold.

02

EVIDENCE WINDOW

The investigation reviews Git activity leading into the selected growth episode.

03

SOURCE LINKS

Findings carry confidence labels and link back to commits and releases.

04

HONEST NULL

When Git history cannot explain a spike, the report says so.

ANALYSIS MODES

Deterministic by default. Optional OpenAI synthesis stays constrained to the collected Git evidence.

OSS InsightGitHub APINext.jsCloudflare
MORE WORK ↓

Explore more