euva-project/euvaA lightweight hex editor and decompiler to solve your binary file analysis problems.
Open platform for reverse engineering.
EUVA isn't just a hex editor and decompiler; it's more of a reverse engineering platform that you customize, rather than having it customize. It's a platform you can plug into anything, and it won't object. We're committed to the community's interests and support maximum tool flexibility and quick integration with other analyzers perhaps your own, or perhaps industrial ones? We don't want to turn the tool into another decompiler, because decompiling isn't fun; IDA, Binija, and others can do it. But we're building something better, trying to build a better architecture, and that's probably our main goal. EUVA will evolve not on its own, but with the community.
This program is under active development. Experimental builds may contain bugs or lead to unexpected behavior. Use with caution.
This software is provided "as is", without warranty of any kind. EUVA is a high-precision instrument designed for educational purposes and security research. The author is not responsible for any system instability, data loss, or legal consequences resulting from the misuse of this tool.
By using EUVA, you acknowledge that you are solely responsible for your actions, you understand the risks of modifying binary files and process memory, and you will respect the laws and regulations of your jurisdiction.
Check out the capabilities of our hex editor, which also includes our DSL for replacing values. This was designed to make it easy to share patches, isn't it? Let's say you're creating something that requires a patch and want to share it with someone else. It's all quite simple: just send them our .euv script, the program applies the patch to the binary file, and you're done no headaches.
Perhaps one of the most important parts of our work is the decompiler it was designed and is being designed with the goal of conveying the meaning of the program's actions as best as possible. We strive to make it user-friendly, because digging through other decompilers is a pain because they act like translators, and you're searching for some kind of meaning in this madness. With our decompiler, the situation is also not simple, but we lower the entry barrier and try to reconstruct the decompiled code as humanly understandable as possible. Of course, we don't do this blindly, and there are limitations to everything, but it's still much better than reading raw C, which isn't even fully translated yet. This is not always convenient, so we are here to try to apply C++ abstractions to decompiled code and be satisfied with the result.
example output:
If you are unsure about the decompiler's arguments, please double-check its results by viewing the disassembler.
If we approach this perfectly, we're not just talking about adding a Themida-like protector detection feature, but also about the decompiler itself. Incidentally, it also supports your patches thanks to C# scripts that are compiled dynamically using Roslyn. These can be called plugins for the decompiler, sort of calibrating it to your needs. This also applies to Yara rules. Perhaps it's the Virustotal integration we recently completed? We don't plan to stop there; we'll continue to develop in this area.
So, we have a dual system: a built-in function in the AI refactoring program where you specify your API key. Points must also comply with the OpenAI standard. This is a quick-intervention button if you don't plan to delve deeply into the process or if you need to conduct analysis discreetly, without the cloud, using local AI models. We also have integration with MCP, which is quite convenient because the model has access to everything it needs. It can read decompiled code and modify it according to your wishes, helping you learn and understand. You can get started with any service that supports MCP by simply configuring the required configuration file. MCP is suitable for those who want to delve deeper and are comfortable using the cloud. It's all your choice. I think we'll keep this flexible approach.
[!NOTE] Your Choice, Your Control: The AI Agent is a "Bring Your Own Key" system. It supports Cloud LLMs (OpenAI, Claude, Groq) and Local LLMs (Ollama, LocalAI). Privacy is paramount.
🟢 After AI (Semantic Refactor)
example output:
