Early access
Welcome to your new favourite reverse engineering toolkit
You're probably already familiar with the normal reverse engineering routine. A whole suite of tools and platforms, each with their own set of supported formats and architectures often hidden behind complex licensing, outputting data in different formats and following different standards.
What if we could bring all of those capabilities together into a single platform? One that doesn't limit the files you can look at based on pricing tiers aimed solely at large companies? Here at Similarity we've decided to do just that, using our years of experience in reverse engineering and malware analysis to build the file analysis platform we've always wanted to use.
Today we're delighted to be able to reveal what we've been working on, and invite you to take part in its future evolution.
Introducing Similarity
Similarity is an all-in-one reverse engineering platform, combining an all-new decompiler with deep data analytics to speed up and simplify your work. Out of the box it will support a wide range of file formats and architectures from standard PE, ELF, and Mach-O binaries to .NET, Java, and scripting languages like Lua and AutoIt. 32-bit, 64-bit, packed, obfuscated - we want the user experience and workflow to be the same regardless. Over time we'll continue expanding support to cover everything you see day to day, guided by uploads and feedback. Upload a file and let Similarity do the work of getting it back as close as possible to the original code so you can get straight into analysis.
See for yourself on these public samples:
| Programs | Languages | Bytecode |
|---|---|---|
| PE | C | .NET |
| ELF | C++ | Android |
| Mach-O | Go | Java |
| APK | Rust | Lua |
| Nim | VB6 P-code | |
| Zig | ||
| Delphi, Pascal | ||
| Objective-C, Swift | ||
| Fortran | ||
| OCaml | ||
| VB6 |
We have built our decompilation from the ground up to solve the challenges and limitations we have faced in reverse engineering. This isn't an existing engine repackaged with new layers on top, Similarity is a whole new approach designed to let us support any file someone might want to analyse, and produce data that can be reliably cross-referenced to help with malware identification, software auditing/verification, version tracking, and any other reverse engineering use case.
On top of decompilation Similarity also acts as an intelligence platform, speeding up file analysis and identification. Known functions - including libraries - will be detected and named automatically as you browse the code, quickly highlighting what's new or modified for further analysis. New versions of existing files can be immediately identified based on function overlap, with the platform identifying matching functions even when changes have been made between releases. As we grow our dataset over time more and more of the files you upload for analysis will be largely identified as soon as you open them. Beyond auto-analysis you can also manually interact with the decompiler and all different aspects of the binary analysis, allowing you to label, edit and customise the output with all the flexibility that’s common across traditional decompilers.
Collaboration is at the heart of the tool. Share your projects with colleagues to give them access to file sets, and work together with realtime viewing and user tracking. Make changes or highlight sections and other users will be able to see it live and make their own additions. Add comments and discuss changes or questions right in the code window with threads and user tagging. All changes are saved to the file's history for easy review so that you can easily keep track of large teams.
Want to bring AI into your workspace? Connect your Agents to the platform, giving them access to the decompiled output and letting you interact with and query them directly from the comments for quick analysis help without ever switching window. They can respond in the platform and propose changes for your review before applying them to the project.
Of course Similarity will also launch with a full HTTP API for scripting support. OpenAPI compatible, it will be available for all platform users to connect scripts and other tooling however you need to. We'll be publishing guides on its usage in the build up to full release.
So what's next?
With this announcement post we are now opening up a waiting list for early access to the platform. The early access period will be invite only based on this list, and will give private access to the current pre-release build of the platform. We expect to start sending those invites out within the next couple of days and weeks and will then continue to expand access until we are ready for a full public release, at which point normal account registration will open.
So far our focus has been on getting the core technology of the platform implemented and working well, but our plans don't end with the release version. With this milestone reached the foundation is in place to start adding additional analysis capabilities on top of the data Similarity will produce, further extending its possibilities for your work.
Those who get early access will have the chance to shape the early evolution of the platform directly, with open feedback channels back to our development team as we make final tweaks and additions. If you have an interest in reverse engineering and think this platform might fit your needs please feel free to sign up at https://simil.ar/register, and we'll reach out with next steps once your account has been approved. Special note: our friends who will gladly wear our merch (t-shirt and polo) will get ahead on the waiting list, don't miss out!
We can't wait to build this platform with you, and to show you what else we have planned. We'll be sharing more content, including early technical details, in the coming days and weeks so make sure to watch this space to find out more. We're also active on various social media where we'll post new updates and future blogposts. Links to those are below.