Recaf FAQs

Question 1

Can Recaf be used for reverse engineering and patching JARs?

Accepted Answer

Yes, Recaf is ideal for reverse engineering. AI assistants can open JARs, decompile code, identify key structures, rename obfuscated symbols, patch bytecode, and export the cleaned or modified JARs, streamlining the entire process.

Question 2

What kind of bytecode operations can AI assistants perform with Recaf?

Accepted Answer

AI assistants can perform extensive operations including decompiling classes, searching code, analyzing call graphs and inheritance hierarchies, direct bytecode modification (renaming symbols, adding/removing methods), compiling Java, assembling JASM, and exporting modified JARs or source code.

Question 3

Which AI clients are compatible with Recaf?

Accepted Answer

Recaf integrates with any AI client that supports the Model Context Protocol (MCP), such as Claude Code and Cursor. It uses a dual-process architecture to bridge the AI client with Recaf's analysis engine.

Question 4

What are the core technical requirements for running Recaf?

Accepted Answer

To run Recaf, you need JDK 22+ and Recaf 4.x. The plugin consists of a Recaf plugin (Bridge Server) and a standalone MCP Server JAR that communicates with your AI client.

Question 5

What is Recaf and how does it leverage AI?

Accepted Answer

Recaf is an AI-powered plugin that enables AI assistants to control Recaf 4.x's robust bytecode analysis, decompilation, and modification capabilities. It integrates directly into AI workflows, allowing natural language commands to manipulate Java bytecode.

Recaf

Recaf

Key Features

Use Cases

Key Features

Use Cases