Should you use it for M.U.G.E.N content?
Ikemen GO supports resources from the M.U.G.E.N engine, but does not promise bug-for-bug emulation. The project README says it aims for compatibility on par with M.U.G.E.N 1.1 Beta and may intentionally differ where matching behavior would preserve bugs or impose limits. Choose it if that policy fits your project, then check the characters, stages and screenpack you intend to use.
Our run on September 30, 2026, at commit 21dd2f2 provides setup evidence, not a gameplay evaluation. We did not test roster compatibility, controller behavior, graphics output, netplay quality or mobile installation. Release v1.0.0 lists rollback and other engine additions, but those release notes do not turn them into features we tested.
What our sandbox actually recorded
Our run recorded installation as successful in 42 seconds. The sandbox used the golang:1.24-bookworm image with 3 CPUs and 8 GB of memory; the installation output shows it downloading Go 1.27.0. The stored dependency count is 77.
The harness records each step's status, exit code, elapsed time and output tail separately. Its build record says exit code 0 after 6 seconds, while the same step's output reports that build constraints excluded the arena package and ends with vet-exit:1. Treat the zero exit status as a recorded command outcome, not proof that a usable engine binary compiled.
The test command returned exit code 1 after 3 seconds. Its stored parser summaries report 0 passed, 3 failed and 3 total. These are parser summaries, not verified counts of unique tests or distinct package failures. The supplied tail names github.com/ikemen-engine/Ikemen-GO/src with [setup failed], reports the arena import error and repeats failure lines under a summary marker. It does not establish three separate failing test cases.
No new execution or complete-log reconciliation accompanied this correction. We did not rerun the command with a different environment. The supplied build-guide extract does not document an arena-specific remedy, so this review does not promise that setting an experiment flag would make the tests pass. No audit result or package-scan coverage is recorded.
The MrKeyoor testing methodology explains the site's approach to command outcomes and sandbox limits. That page supplies provenance context; its current procedure does not establish which methodology version governed this older run.
Release archives and source builds answer different needs
The README provides ready-to-use release archives. Those are the direct starting point for someone evaluating M.U.G.E.N content. The current source-build guide is for engine development and requires Go 1.27 or newer plus native media libraries and build tools. Its helper script can build minimal FFmpeg rather than requiring system FFmpeg development packages.
A desktop executable also needs the runtime asset folders described in that guide. Compilation and running a complete game are separate checks. Our sandbox record establishes neither a successful launch nor a working asset bundle.
The current README lists Windows 10, macOS 13 on Apple Silicon, Linux with glibc 2.35 or newer, Android 13 and iOS 17 as minimum platforms. It also requires OpenGL 3.3, OpenGL ES 3.2 or Vulkan. The README at our recorded commit instead lists Android 14, so use the requirements attached to the version you download rather than applying today's table to the old run. Published Intel macOS builds are not currently available, although the current build guide describes building for Intel Macs.
Android releases use an APK. iOS uses a sideloaded IPA, and the README says free Apple IDs require refreshing the app every 7 days. Our sandbox did not validate either installation path.
When to consider another engine
Castagne is a layer built on Godot that manages fighting-game logic. Its README makes it a relevant candidate when you want to build a fighter around Godot. The supplied evidence does not establish M.U.G.E.N resource compatibility, so do not assume it can replace Ikemen GO for an existing roster.
Godot provides a unified interface for creating 2D and 3D games. Its README supports considering it when your project needs a general game engine. Godot is also Castagne's foundation, so these two choices overlap. Neither supplied extract establishes a direct M.U.G.E.N migration path.
Distribution and rating limits
The engine source is MIT licensed, bundled screenpack assets have Creative Commons terms, and FFmpeg has LGPL v2.1 obligations. Check each component and any added assets before distributing a game bundle.
Scores use a five-point editorial rubric: 1 means substantial documented obstacles, 2 means major unresolved limitations, 3 means usable with material caveats, 4 means substantial supporting documentation or project evidence with remaining limits, and 5 requires broad verification for the dimension being rated. These are judgments, not benchmark results. Setup considers command outcomes and prerequisites; documentation considers the supplied setup and licensing coverage; community considers visible contribution evidence; maturity considers the release record. Community does not measure adoption or support response time, and maturity does not certify gameplay reliability.

