Smarter Prioritisation with Auto-Deep-Analysis

Release v0.1.72 · Top-N Auto-Deep-Analysis Selector

---

One of the platform's core promises is that you paste a directory URL and it surfaces the most buildable, high-demand products — without you having to manually read every listing. With v0.1.72, that promise takes a meaningful step forward.

The Problem We Solved

A single proptech directory like Kerfuffle can list hundreds of products. A full deep-analysis of every listing is slow and expensive. But deciding which products deserve deep analysis by hand defeats the point of an automated intelligence tool.

What We Built

After a skim crawl completes, the platform now automatically scores every product on two surface signals — review count (a proxy for market traction) and pricing tier (a proxy for revenue opportunity) — and flags the top N as candidates for deep analysis. N defaults to 20, but it's configurable.

Critically, the system doesn't fire the deep-scrape pipeline immediately. It pauses and presents you with the flagged list first. You can adjust it — add a product you spotted, drop one that's out of scope — and then confirm. Only then does the deep analysis run.

Why a Human Checkpoint?

Surface signals are a good first filter, but they're not perfect. A product with few reviews might be new and fast-growing. A free-tier product might be building toward a paid model. The confirm step costs you thirty seconds and gives you full control over what gets analysed deeply.

What This Enables

This release directly implements the workflow described in our platform mission:

"Core workflow: paste a directory URL, system skims all listings, deep-analyses the top scoring products, presents ranked results."

With v0.1.72, the bridge between skim and deep-analysis is now automated and production-ready.

---

See the Top-N Auto-Deep-Analysis Selector feature guide for full configuration details.