Case study · Live
Opportunity Intelligence
A local-first market intelligence engine that reads public business conversations, scores the opportunities hiding in them, and publishes its own research every week, without a single paid API call.
The problem
Builders guess. Then they build the wrong thing.
Public communities are full of people describing expensive, repeated problems in their own words. That signal is buried in noise, scattered across platforms, and gone by next week. Most founders never see it. They build from intuition and validate after the fact, when changing course is most expensive.
The approach
A pipeline instead of a hunch
Five deterministic stages, each one inspectable and reproducible. The same run always produces the same verdict.
- 01
Collect
Pulls public posts from Reddit business communities (RSS) and Hacker News (Firebase API). No scraping tricks, no paid data.
- 02
Detect pain
Deterministic keyword scoring finds posts where people describe real problems, weighted toward what they put in the title.
- 03
Score opportunities
Each candidate gets a 0 to 100 score across pain frequency, community spread, monetization, build difficulty, competition, and cross-source validation.
- 04
Find gaps
Opportunities are grouped into industry-level gaps, so a single loud post never outweighs a quiet repeated pattern.
- 05
Publish
A weekly digest becomes an article, a newsletter, and a publish-ready package. Exported as JSON straight into this website.
Live metrics
Straight from the pipeline
These numbers are exported by the engine itself on every run. The same JSON this page is built from.
- 510
- Public signals analyzed
- 150
- Pain signals detected
- 11
- Opportunity candidates
- 7
- Industry gaps found
- 11
- Opportunities tracked
Last export: June 10, 2026
See it work
The system publishes its own proof
Or go back to all projects.