Platform
ls-judge: 65+ languages, each in its own microVM
ls-judge is the public code-execution API that backs Omni Compiler, the Online Code Editor and Code Sprout's practice problems. 65+ languages, 120+ runtimes, every submission in its own Firecracker microVM.
Who it's for
Developers using Omni Compiler or the Online Code Editor, teachers assigning programming exercises, and anyone poking at the public endpoint to run a snippet quickly.
What it does
A public code-execution API: POST a submission (source, language, optional test cases, time + memory limits), get a verdict back per test. 65+ languages, 120+ runtimes (multiple versions of C++, Python, Node, Rust, Go, Java, Kotlin, Swift, Julia, Raku, Zig, and more). Streams results as each test finishes so a long batch does not block the UI.
How we built it
Every single run executes in a fresh Firecracker microVM — not a Docker container, not a chroot, a full virtualisation boundary — with a time budget, memory ceiling and no outbound network. The VM boots in under 100 ms thanks to pre-warmed snapshots, so throughput is high despite the isolation. For deeper detail, see the follow-up ls-judge engine tour.
Open: runtimes.laxmi.solutions.