Can You Understand How an App Feature Works Without Its Source Code?

Can you understand how a feature in any app works without looking at its source code? With REA it has become much easier. The project has passed 150k stars on GitHub.

REA is, at its core, an MCP server: one program that puts many reverse engineering tools in a single place. You connect to it from an agent like Codex, Claude Code or Cursor and ask how the feature you are curious about works.

REA returns its results with evidence: what was observed, what was inferred, what is unknown. The agent interprets them and reports back to you.

One of its biggest advantages is how many target types it supports: native binaries (through Hopper, Ghidra or IDA), JavaScript and Electron apps, .NET assemblies and Android APKs.

Why use REA when a normal agent could do the same job?

It could. REA adds three things.

First, it puts many engines under one contract. Decompiling works with the same interface in Hopper and in Ghidra. Targets that normally each need a separate tool are in one place.

Second, it returns results with evidence, which makes it harder for the agent to make things up.

Third, it takes over operational work like starting the engine, timeouts and caching, so the agent doesn't have to deal with it.

So what do we do about it? The other day I came across a video by Ali Kutluözen that talks about exactly this. In his words (translated from Turkish): "Do you have a small app idea? Did you spend months working on it and build a product? Now people can copy the same thing in minutes."

I recommend watching it: https://www.youtube.com/watch?v=J3ws3sf_iqM

Note: REA is for legal use only, such as your own software, interoperability and research. Copying someone else's product without permission is still a legal problem.

GitHub: https://github.com/morluto/rea