AI Can Copy Any App Now - Here's What People Will Still Pay For

Duncan Rogoff | Learn Claude Code

Key takeaways

People still pay for measurable results, dependable support when software breaks, and the trust and audience of the people behind a product. The cost of copying software is falling, but the copies may not yet match the original for professional use.

REA can help inspect apps without their source code, making it useful for security work, learning, and preserving old software, but copying and selling someone else’s paid product raises ethical and legal risks. To test a product’s moat, ask why customers would stay if it were copied, then assess its results, support, and existing customer trust.

Video Flow

  1. 0100:00

    REA’s sudden surge and the moat question

    REA, a repository for reverse engineering apps, rose from about 13,000 to 149,000 GitHub stars in five days as interest in AI-made app copies grew. A practical moat test is to ask why customers would stay if a product were perfectly copied; reasons such as fast responses, customer knowledge, and a good working relationship do not depend on code.

  2. 0202:12

    What REA does and how to try it

    REA is a free add-on that works with Claude Code, Codex, Cursor, and Gemini CLI, using MCP to let an agent direct reverse-engineering tools such as Ghidra, Hopper, and IDA. The agent can investigate an app, point to evidence for its findings, report what it could not determine, and use that analysis to build a similar feature; screenshots or access to the app can help it inspect the target.

  3. 0304:30

    Examples of reverse-engineering results

    REA’s inspection of the PowerToys color picker produced eight fully correct findings out of nine, which were used to build a new picker in under 750 lines; a DX Ball reconstruction matched the original in 3,205 tests down to the byte. A developer also used a QuickBooks trial to study the accounting app and build an open-source alternative, while acknowledging that REA did not provide the original code or a finished product in one step.

  4. 0406:10

    Appropriate uses and limits

    REA is intended for security research, learning, interoperability, software preservation, and checking software one owns; its rules reject breaking copy protection, making cheats, and copying paid software. Reverse engineering or selling a copy may also breach an app’s license depending on location and agreement, so it suits security practitioners, people rescuing abandoned software, and learners more than beginners.

  5. 0508:05

    Outside-in copying versus inspecting the internals

    Art Craft rebuilt seven Adobe apps, including Photoshop, Premiere, and Lightroom, from scratch by studying public documents and app behavior rather than Adobe’s code; its Photoshop copy was not yet ready for daily professional work. REA instead inspects an app’s internals, so outside-in reconstruction and inside-the-box analysis lower copying costs in different ways without establishing that the resulting copies equal the originals.

  6. 0609:12

    Customers pay for outcomes, support, and trust

    A product that books sales calls can command more than a tool that only drafts posts because customers pay for the result and for someone who knows what works and stands behind it. Reliable support matters when a free accounting alternative fails at a critical moment, while an established audience or community helps a product find buyers among many competing apps.

  7. 0712:23

    Turn the moat into a practical check

    Software is not dead, but ease of building an app is becoming less valuable relative to outcomes, support, trust, and the ability to reach customers. A product check asks whether customers can name a real result, know how to get help when something breaks, and already know or trust the people behind it; the weakest answer identifies what to improve.

Tools & Resources

A repository that gives AI agents tools to inspect apps without their source code.

  1. 01
    A repository that gives AI agents tools to inspect apps without their source code.00:04
  2. 02
    A repository that gives AI agents tools to inspect apps without their source code.02:29
  3. 03
    A repository that gives AI agents tools to inspect apps without their source code.06:49

An AI coding agent that can install and operate REA.

  1. 01
    An AI coding agent that can install and operate REA.02:42

A coding agent listed as compatible with REA.

  1. 01
    A coding agent listed as compatible with REA.02:42

An AI coding agent listed as compatible with REA.

  1. 01
    An AI coding agent listed as compatible with REA.02:42

A command-line AI agent listed as compatible with REA.

  1. 01
    A command-line AI agent listed as compatible with REA.02:42

The protocol REA uses to provide tools to an AI agent.

  1. 01
    The protocol REA uses to provide tools to an AI agent.02:47

A reverse-engineering program that REA can drive to inspect app code.

  1. 01
    A reverse-engineering program that REA can drive to inspect app code.03:26
  2. 02
    A reverse-engineering program that REA can drive to inspect app code.07:48

A reverse-engineering program that REA can drive to inspect app code.

  1. 01
    A reverse-engineering program that REA can drive to inspect app code.03:26
  2. 02
    A reverse-engineering program that REA can drive to inspect app code.07:48

A reverse-engineering program that REA can drive to inspect app code.

  1. 01
    A reverse-engineering program that REA can drive to inspect app code.03:26

Asks an agent to investigate search in a notes app, provide evidence, and build a similar feature.

  1. 01
    Asks an agent to investigate search in a notes app, provide evidence, and build a similar feature.03:58

The app whose color picker was inspected as a reverse-engineering test case.

  1. 01
    The app whose color picker was inspected as a reverse-engineering test case.04:43

The 1990s brick-breaker game used to test reconstruction of game behavior.

  1. 01
    The 1990s brick-breaker game used to test reconstruction of game behavior.05:06

The accounting app used as the subject of a reverse-engineering example.

  1. 01
    The accounting app used as the subject of a reverse-engineering example.05:36
  2. 02
    The accounting app used as the subject of a reverse-engineering example.11:13

A project that rebuilt Adobe apps from the outside using public documents and observed behavior.

  1. 01
    A project that rebuilt Adobe apps from the outside using public documents and observed behavior.08:10

An Adobe app whose open-source copy was cited as not ready for daily professional work.

  1. 01
    An Adobe app whose open-source copy was cited as not ready for daily professional work.08:29

An Adobe app included among the outside-in copies.

  1. 01
    An Adobe app included among the outside-in copies.08:31

An Adobe app included among the outside-in copies.

  1. 01
    An Adobe app included among the outside-in copies.08:31

A paid learning community used as an example of customers paying for teaching, structure, and trust rather than freely available content alone.

  1. 01
    A paid learning community used as an example of customers paying for teaching, structure, and trust rather than freely available content alone.09:32
  2. 02
    A paid learning community used as an example of customers paying for teaching, structure, and trust rather than freely available content alone.13:36

Share a little
cookie?

Your cookies help us understand visits and how you use Vigest, so we can make it better. We use Google Analytics. You can use the service without agreeing.

You can change your choice anytime.
Saved in this browser for 180 days.

Privacy notice