ShotForge

ShotForge is a local-first iOS app for basketball training that uses computer vision to detect shots and analyze mechanics. The description states it was built as a hackathon project by one developer (Blair Ding) using Swift, SwiftUI, and Apple’s Vision framework. It claims to track performance and analyze mechanics using iPhone cameras, with optional Apple Watch support. The author describes building an offline evaluation system and training models for ball detection and pose estimation.

The single most important open question is: What is the actual commercial viability of this product as a standalone tool for basketball players?

The description contains no evidence of revenue, customers, or adoption beyond self-reported development metrics. It does not state whether the app has been released to users, how many shots it can process, or what its accuracy means in real-world use. The author makes claims about technical capabilities but provides no data on user behavior, retention, or monetization.

This is a self-reported project description with no external verification. All findings are based solely on the author’s own account and must be treated as unverified claims.

The description states that ShotForge is:

• A local-first iOS basketball training app
• Built using Swift, SwiftUI, AVFoundation, SwiftData, and Apple Vision framework
• Designed to track performance (makes, misses, court locations, shot types) and analyze mechanics (body movement such as knee flexion, elbow angle, release height)
• Uses a face-on camera view for mechanical analysis
• Processes video through multiple stages including ball detection, tracking, and result classification
• Stores data locally by default with no account required

It also states that it was built as a hackathon project submitted to the OpenAI 2026 hackathon.

Inferred from the description:

• The app is designed for personal use in basketball training
• It includes both performance tracking and mechanics analysis modules
• It uses on-device processing without cloud dependencies
• It allows uncertain results to be left unconfirmed rather than forcing a classification

Not evidenced:

• Whether the app has been released or made available to users
• If it supports real-time feedback or integration with other platforms
• How many shots it can process per session
• Any pricing model or monetization strategy

The description states that ShotForge aims to:

• Bring together performance tracking and mechanics analysis in one tool
• Help players understand their own movement, measure consistency, and connect changes in technique with performance outcomes
• Avoid requiring coaches, specialized equipment, or manual video review
• Not force players into a single definition of “perfect form”
• Use hardware already owned by players (iPhone, optionally Apple Watch)

It positions itself as an alternative to tools that focus only on results or only on mechanics.

Inferred from the description:

• The product is positioned for individual basketball players seeking self-improvement
• It emphasizes personalization and autonomy over external coaching
• It targets a niche market of tech-savvy athletes who prefer local processing

Not evidenced:

• Whether this positioning has been validated with actual users
• How it differentiates from existing apps or tools in the space
• Any marketing claims or user testimonials
• Evidence of product-market fit beyond developer experimentation

The description states that ShotForge is intended for:

• Basketball players who want to improve their shooting skills
• Users who prefer using hardware they already own (iPhone, optionally Apple Watch)
• Players looking for feedback without a coach or expensive equipment

Inferred from the description:

• The primary customer segment appears to be amateur and recreational basketball players
• It may appeal to those interested in data-driven training methods
• The optional Apple Watch support suggests targeting users with wearable tech access

Not evidenced:

• Specific demographics or user personas
• Customer acquisition channels or marketing strategies
• Any evidence of actual users or target audience testing
• Whether the app targets professional or youth players

The description states:

• No account is required
• Data stays on device by default
• The app uses local processing and does not rely on cloud services
• It was built as a hackathon project

Inferred from the description:

• The business model likely involves no recurring revenue or subscription fees
• It may be free to download with optional premium features (not stated)
• There is no mention of monetization strategy beyond the app itself

Not evidenced:

• Any pricing structure, subscriptions, in-app purchases, or monetization plans
• Revenue streams or business model assumptions
• Whether there are plans for paid upgrades or enterprise versions
• Customer lifetime value or monetization roadmap

The description states:

• Built as a native Swift and SwiftUI application
• Uses AVFoundation for video capture, SwiftData for sessions, Apple Vision framework for body pose extraction
• Video pipeline includes stages: detect player/ball/rim → track ball → identify shooting events → classify result (made/missed/unknown)
• Trained an RF-DETR-based model for small-ball detection
• Uses temporal tracking and physically meaningful trajectory rules
• Includes a separate offline evaluation system for experimenting with models
• Apple Watch prototype captures high-frequency wrist motion but is optional

Inferred from the description:

• The app leverages Apple’s native frameworks, suggesting compatibility with iOS ecosystem
• It uses machine learning models trained on basketball footage
• It handles uncertainty in detection by leaving results unconfirmed when ambiguous
• It separates development and production pipelines for model testing

Not evidenced:

• Performance benchmarks beyond development metrics
• Real-time processing capabilities or latency data
• Scalability of the system or handling of large volumes of data
• Integration with external devices or platforms
• Any deployment or distribution infrastructure

The description states:

• Development results: 90.1% precision and 94.4% recall for shot detection
• Make-or-miss accuracy of 87.6% with 82.9% coverage on a limited dataset
• Correctly classified all eight short validation clips
• These are described as development results, not universal claims

Inferred from the description:

• The project has reached a functional prototype stage
• It includes reproducible datasets and manual auditing of labels
• It demonstrates engineering discipline in evaluation and testing

Not evidenced:

• Any real-world usage or adoption data
• Customer feedback or user engagement metrics
• Product release status or availability to end users
• Growth trends, retention rates, or conversion data
• Evidence of product-market fit or market traction beyond the developer’s own testing

The description does not provide any information about:

• Competitors in the basketball training or analytics space
• How ShotForge compares to existing tools or platforms
• Market size or competitive landscape analysis
• Any differentiation strategies or unique value propositions relative to competitors

Not evidenced:

• Competitive positioning or market share data
• Pricing comparisons with similar tools
• Feature sets or capabilities of competing products
• Industry trends or adoption patterns in basketball training tech

Based on the description, several risks and red flags are evident:

1. No commercial traction: The app is described as a hackathon project with no evidence of user base, revenue, or market validation.
2. Limited dataset: All performance metrics are based on a limited set of development videos under controlled conditions.
3. Unclear monetization: No indication of how the product will generate revenue or sustain itself beyond its initial development phase.
4. Single developer team: With only one member (Blair Ding), scalability and long-term maintenance may be concerns.
5. Technical limitations: The app relies heavily on local processing, which may limit real-time performance or advanced features.
6. Ambiguity in user experience: While it claims to handle uncertainty well, there is no clarity on how this affects usability or adoption.

Not evidenced:

• Any risk mitigation strategies or contingency plans
• Evidence of market demand or competitive pressure
• Long-term roadmap or product evolution beyond the current prototype

1. Has the app been released to users? If so, how many active users are there?
2. What is the actual user feedback on accuracy and usability in real-world settings?
3. Are there any plans for monetization or revenue generation beyond the initial prototype?
4. How does the system perform under varying lighting conditions, court types, and player movements?
5. What are the technical limitations of the current implementation that might affect scalability?
6. Is there a plan to expand beyond iOS or integrate with wearable devices like Apple Watch more deeply?
7. Have you tested the app with actual basketball players, not just developers or internal testers?
8. How do you intend to validate performance metrics in real-world usage?