Skip to content
OpenSourceSoftware.ai
REA logo

REA

morluto · Developer platforms

GitHub stars

385,615
—Collecting 30d data

Stats updated

151,941 stars on Oct 12 to 385,615 stars on Oct 26, up 233,674.15 measured star snapshots

What is REA?

Start with a question about software you own or have permission to inspect. REA helps your coding agent gather the implementation details needed to answer it, turning an investigation into a sequence of evidence-backed steps rather than a guess based on the interface alone.

REA connects coding agents to local tools for inspecting native binaries, JavaScript and Electron apps, .NET assemblies, and websites. It returns code, call relationships, and source evidence to help explain existing behavior or rebuild a feature. Use it through MCP or the command line; native analysis needs a compatible engine such as Ghidra, Hopper, or IDA.

Core capabilities

  • Trace features through packaged apps, native binaries, and browser activity.
  • Connect local analysis tools to your coding agent through MCP or use the CLI.
  • Review code locations and supporting evidence alongside each finding.

Follow a feature through an application

For JavaScript and Electron projects, REA maps modules, imports, routes, and communication between processes. It can examine extracted application folders and ASAR archives. Browser workflows add page structure, loaded scripts, and network observations, helping an agent connect a visible interaction to the code that implements it.

  • JavaScript
  • Electron
  • Browser analysis

Bring native analysis into your agent workflow

REA connects to analysis engines such as Ghidra, Hopper, or IDA to inspect functions, assembly, strings, and references. Setup registers a local MCP server with a supported coding agent; terminal commands expose the same workflows. Supported operating systems and formats vary by engine, and commercial tools require their own licenses.

  • Binary analysis
  • MCP
  • CLI

Check the evidence and understand the limits

Findings include evidence and limitations so you can check an explanation against the original program. Static inspection reads files, while runtime observation may launch or interact with the target. Analysis runs locally, but your coding agent's model provider may receive tool results under its own data policy. Use REA only for software you are authorized to examine.

  • Source evidence
  • Local analysis
  • Verification

Where it fits

Use cases

  1. 01

    Understand a feature in an existing app

    Ask how an export, clipboard action, or calculation works. Follow the relevant modules and calls, then use the findings to document the behavior or plan an interoperable implementation.

  2. 02

    Recover a small routine from a binary

    Inspect an executable you are authorized to analyze. Use disassembly, pseudocode, and references to reconstruct a specific calculation, then compare the new implementation against the original behavior.

  3. 03

    Investigate browser behavior

    Connect to a browser session and inspect the scripts and requests behind a page interaction. The official dinosaur-game example shows how a recovered speed rule can inform a small reconstruction.