What it does
When software is built, its human-readable code is turned into machine code. Decompiling tries to reverse this. LLM4Decompile is a family of AI models trained to do that for Linux programs, producing C code.
Some versions work straight from the machine code; others improve the rough output of an existing tool called Ghidra. The models come in several sizes and are shared on Hugging Face.
The project also publishes test sets and training data so others can measure and build on the results.
Who it's for
Researchers and engineers working on reverse engineering and program analysis.
Can I use it without coding?
Developers only. Needs Python setup and technical knowledge; a demo notebook is offered. New to this? Start here.
Can I run it?
- Cost
- FreeNothing to pay or sign up for.
- Runs on
- Linux
- Computer
- Works best on a powerful computer or one with a graphics card (GPU)
- Privacy
- Runs on your own device: your data stays with you
- API key
- Not needed
Try it
Using Claude Code, Cursor or another AI coding assistant? Paste the prompt and it will walk you through the setup.New to this?
Before you install
- Only download from the official GitHub page. Copies with similar names exist for many popular projects.
- License (MIT): Very permissive: free to use, change and share, including for business.
- More on staying safe
Where people are talking about it
- LLM4Decompile: Decompiling Binary Code with LLMHacker News · Mar 2024
- LLM4Decompile: Decompiling Binary Code with Large Language ModelsResearch paper · Mar 2024