Learning and careers decision
Swish AI
A technical user can build a narrow self-hosted shot-tracking and pose-analysis workflow, but reproducing the polished model quality, continuous ML improvements, and full mobile polish of the paid app is unlikely without the vendor's proprietary work.
View on the App Store↗$5.99/mo
$72/yr
Read off the official pricing page.
$100one-off64 h to build
$250/mo6 h/mo upkeep
On cash alone, building overtakes the subscription at 44 seats.
No open-source build does this yet
Nothing published replaces this one, so a replacement starts from an empty file. Here is what it would have to cover.
What a replacement has to do
- User records a shot session on iPhone → upload or process video → run pose/shot detection per frame → compute form metrics and generate feedback → show metrics and drills in mobile UI
What it still won’t have
- Proprietary model quality and ongoing ML improvements
- Polished App Store UX, ratings, and discoverability
- Continuous mobile platform bugfixes and performance tuning
- Any backend telemetry and analytics tied to 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 44 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 Swish‑like iOS app using Swift (UIKit or SwiftUI) and a small server. Stack: iOS Swift app, Node.js + Express backend, PostgreSQL for metadata, S3-compatible object storage, and a GPU-capable inference endpoint (FastAPI or Flask) running an open-source pose estimator. In‑scope: iOS camera UI to record/trim video and upload, server endpoints to receive videos and store metadata, a worker to extract frames and run pose estimation, a rules-based analyzer that computes elbow alignment, foot placement, release timing, follow-through and a simple Form score, mobile screens to display per-shot metrics, history and a basic training-plan generator, and StoreKit subscription integration with entitlement checks. Out of scope: training new ML models from scratch, multi-user analytics dashboards, advanced drill recommendation engines, and App Store submission support. Include error handling, retries for uploads, basic unit tests for server logic, and end-to-end integration tests for the upload→analysis→display flow.
How we checked
How the score was reached
- Partly verdict base52
- 2 cited sources+1
- Price verified on pricing page+3
- 3/3 assessment runs agreed+4
- Evidence score60
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 · 2
Every page the run actually retrieved.
- official productSwish AI: Basketball Coach App - App Store
- official pricingSwish AI: Basketball Coach App - App Store (pricing section)
Integrity checks
What held up, and what did not.
