# Demand Radar

> Demand Radar scores product demand from 0 to 100 for any keyword or niche, using cited public data instead of generated guesses. Sixty points are computed from Hacker News, Stack Overflow, the App Store and GitHub. Forty points are assigned by the user through a fixed checklist about payment evidence and buildability.

Demand Radar exists because idea lists are unverifiable. Every number in a scan can be traced back to a public source, and the raw complaint text is shown verbatim next to the score.

## How to use it

Open https://demandradar.pages.dev/ and enter a niche, for example "invoice automation" or "hoa dues tracking". The scanner returns a verdict (build it / worth watching / let it go), a six-part score breakdown, and the review and forum text that produced it.

## The rubric

- Demand volume, 20 points, automatic: Hacker News comments in the last 12 months, plus Stack Overflow questions at 0.6 weight, plus Reddit threads at 1.5 weight, log-scaled.
- Complaint intensity, 20 points, automatic: share of up to 100 Hacker News comments and 100 Reddit threads containing pain language, blended with absolute count.
- Competition gap, 20 points, automatic: number of US App Store apps with 500+ ratings whose title matches the keyword, with GitHub repository count at one twentieth weight.
- Payment evidence, 20 points, user-assigned: people already paying for a workaround; a worse tool already charging.
- Buildability, 15 points, user-assigned: no license required, two-week core journey, no inventory or human ops.
- Reuse, 5 points, user-assigned.

Verdicts: 70 and above is build it, 50 to 69 is worth watching, below 50 is let it go.

## Pages

- https://demandradar.pages.dev/ : the scanner
- https://demandradar.pages.dev/method.html : the full rubric, data sources and known limits
- https://demandradar.pages.dev/pricing.html : plans and current billing status

## Data sources

- Hacker News via the Algolia search API: comment volume (approximate, as Algolia reports it), pain-language analysis, verbatim quotes.
- Stack Overflow via the Stack Exchange API: question volume.
- Reddit via /api/reddit, a Cloudflare Pages Function that authenticates with OAuth client_credentials and calls oauth.reddit.com/search, falling back to anonymous JSON. Edge-cached 30 minutes. Reddit blocks direct browser requests and most datacenter IPs.
- App Store via /api/itunes, a Cloudflare Pages Function fronting the iTunes search API (cached 1 hour) and the customer reviews RSS (cached 30 minutes). Apple throttles the review feed by IP; the proxy detects the resulting empty feed and reports it as an error rather than as "no reviews".
- GitHub via the search API: repository density.

## Known limits

- Reddit access depends on the proxy and on Reddit not blocking the request's IP range; when it fails the scan marks the source unavailable rather than scoring zero.
- Hacker News and Stack Overflow are developer-biased, so non-technical niches score artificially low on volume.
- App Store data is US-only.
- The score is a snapshot, not a forecast.

## Machine-readable

- https://demandradar.pages.dev/ai-catalog.json
- https://demandradar.pages.dev/sitemap.xml

## Contact

Method questions and source corrections: open an issue on the repository linked from the site footer.