CardScope

What the company appears to be: CardScope is a self-reported cross-platform mobile application designed to scan Pokémon cards in real time using on-device machine learning (ML), with the goal of quickly identifying card values and enabling efficient trading at card shows. The app uses a distilled version of a larger ML model (Google SigLip) trained to run locally on smartphones, avoiding reliance on internet connectivity or cloud processing.

What changed: The project was built as a hackathon submission for the OpenAI 2026 hackathon. It is described as a proof-of-concept with no evidence of commercial traction, revenue, or user adoption. The team claims to have developed an ML model that can run on-device and perform card recognition and value lookup in real time.

Single most important open question: Is there any evidence that the app has been released to users beyond the founders' local network, or that it has achieved meaningful usage or feedback from external users?

The description states that CardScope is a mobile application that:

• Scans Pokémon cards in real time using on-device machine learning.
• Detects card bounding boxes and crops them for processing.
• Uses an embedding model to search against a database of Japanese and English Pokémon cards.
• Provides pricing information from major sites.
• Allows users to track collections, make sub-collections, and view total value.

It is described as running fully on-device to work even with poor internet connectivity, such as at card shows.

Evidence: Self-reported by the authors. No independent verification or demonstration provided.

The project description states that CardScope was built to solve problems faced by Pokémon card traders:

• Manual price lookups are tedious.
• Existing apps only support US markets.
• Single-card scanning is inefficient.
• Cloud-based scanning fails in areas with bad internet.

It positions itself as a solution for fast, cross-platform, on-device card scanning and value lookup.

Inference: The positioning reflects the authors' personal experience rather than market research or user validation. It is not clear whether this addresses a broader market need beyond their own use case.

The description states that the app targets Pokémon card traders who attend local leagues and card shows, where internet connectivity may be unreliable.

Evidence: Self-reported by the authors. No data on customer segments or personas is provided.

There is no evidence in the description of a business model or pricing strategy. The project is described as a hackathon submission with no indication of monetization plans, subscriptions, or paid features.

Evidence: Not evidenced.

The app is built using:

• Cross-platform tools (Expo.io)
• On-device ML models (RepViT, distilled from Google SigLip)
• Mobile development frameworks (Swift, Android)
• AI experimentation via Codex and GPT-5.6 Sol Pro
• Backend components like Drizzle, Turso, CoreML

The team reports that they used a teacher-student model approach to optimize performance for mobile devices.

Evidence: Self-reported by the authors. No independent technical review or product demonstration is available.

There is no evidence of any traction, revenue, customers, or user adoption beyond the founders' local testing and feedback from friends at a Pokémon League.

The project was submitted to a hackathon and has not yet been released on public app stores.

Evidence: Not evidenced.

The description mentions that existing apps do not support European markets and only allow single-card scanning. However, there is no mention of competitors or market analysis beyond these claims.

Evidence: Self-reported by the authors. No competitive landscape data provided.

• Unverified claims: All descriptions are self-reported and unverified.
• No commercial traction: No evidence of users, revenue, or adoption.
• Limited scope: The app appears to be a prototype built for a hackathon with no indication of scalability or long-term viability.
• Founder dependency: Only one team member (Gaden) is mentioned; no additional team structure or roles are described.
• Unclear monetization path: No evidence of how the product will generate revenue.

Inference: The lack of any external validation, user feedback, or commercial activity raises concerns about whether this represents a viable business opportunity.

1. Has the app been released to users outside of your local network?
2. What is the current state of the on-device ML model in terms of accuracy and performance?
3. Are there any plans for monetization or user acquisition beyond initial testing?
4. How do you plan to scale the card database and pricing data sources?
5. Have you considered how to handle edge cases like low-quality scans, damaged cards, or non-Pokémon cards?