Image and video decision
Watch Identifier
A practical, limited replacement is realistic for a single capable developer using existing open-source vision libraries and a small hosted inference endpoint; the vendor’s proprietary training data and curated pricing are the main non-trivial losses.
View on the App Store↗$1.67/mo
$20/yr
Read off the official pricing page.
$100one-off60 h to build
$30/mo6 h/mo upkeep
On cash alone, building overtakes the subscription at 23 seats.
Open-source builds that already do this
Every project below is open source and already does this job today. Fork one, self-host it, or take the parts you need - the build prompt further down assumes an empty file, and this is the shortcut past that. Licences differ; check the one on each card before you ship. All Watch Identifier alternatives, with the arithmetic →
What a replacement has to do
- Upload or capture watch image → run image classifier / feature extractor → map to brand/model and pricing table → run authenticity heuristics → return results and save scan history
What it still won’t have
- Proprietary training data and any vendor-tuned authenticity model
- Curated market-price dataset and trend insights
- App Store userbase, ratings, and established UX polish
- Any server-side scoring optimizations and on-device model optimizations used by the vendor
What remains hard
- Product polish and ongoing maintenance
First-year cost
Keep paying
Paying is—cheaper in year one.
On cash alone, building overtakes the subscription at 23 seats.
Money you would actually spend
Time you would spend
—
What you would spend
What we assumed
The verdict above measures whether you could build it. This one is only about money.
Runnable build prompt
Build a minimal Watch Identifier iOS app using SwiftUI + Combine and a Python FastAPI backend (host on a $10/mo VPS or serverless) that: 1) accepts camera or gallery images and runs client-side preprocessing (crop to watch, resize, normalize); 2) sends images to a hosted inference endpoint (TorchScript/CoreML or a simple REST inference server) that returns brand/model and feature embeddings; 3) matches predictions to a lightweight Postgres (or Firestore) table of known models and prices; 4) runs a set of deterministic authenticity heuristics comparing embeddings and key landmark positions; 5) stores scans and shows a scan history and basic trend/price field. Out of scope: building a production-grade proprietary dataset, large-scale pricing analytics, or advanced forensic authenticity models. Include error handling for network and inference failures, basic unit and integration tests for the backend endpoints and model integration, and CI to run tests before deployment.
How we checked
How the score was reached
- Build verdict base78
- An open-source build was found+5
- 4 cited sources+3
- Price verified on pricing page+3
- Evidence score89
The base comes from the verdict. Everything under it is a check that either happened or did not, and each one is a fact frozen in this record rather than a judgement made at render time - so the same evidence always produces the same number.
How scoring works →Cited sources · 4
Every page the run actually retrieved.
- official productHoro ID - Watch Identifier - App Store
- official pricingHoro ID - Watch Identifier - App Store (pricing)
- open sourceOlafenwaMoses/ImageAI
- open sourcePaddlePaddle/PaddleClas
Integrity checks
What held up, and what did not.
