Source: erkamyaman/rust-anything (MIT, created 2026-11-16), surfaced by André Lindenberg on LinkedIn. Distributed on crates.io and npm; registers itself as an MCP server and skill for Claude Code, Codex, Cursor and Gemini CLI.
The idea
rust-anything does not translate code. The coding agent (or a person) writes the Rust; the tool supplies what an agent cannot reliably guess: where the time goes, what to port first, a bridge so existing callers keep the original API, the exact behavior the port must match, and a gate that defines done. The original program stays the source of truth throughout.
It supports TypeScript/JavaScript, Python, C, Go, Ruby and Java, each with a native bridge (napi-rs or wasm, PyO3, extern "C", cgo, magnus, JNI) and a driver that records and replays the original.
The loop
Profile a real workload with the language's profiler to find hot functions and loops worth batching; it flags code that is already compiled, where Rust gains little.
Analyze units, port order, risks with file and line, and dependency-to-crate fit.
Bridge generates a crate with bindings and
port.rsstubs;--batchkeeps many values inside Rust so a call crosses the boundary once.Record runs the original on generated edge cases, built-in traps and user inputs, saving outputs and errors.
Port (agent or human writes
port.rs).Diff replays every recorded case against the Rust.
Bench gives a faster-or-not verdict; calling Rust once per small function is often slower than a JIT, so the tool aims at loops.
Shadow runs both side by side in production.
Gate must pass before a unit counts as done.
For a whole app, the same loop runs unit by unit inside an outside-in frame: record the app as a black box from its documented entry points, let a Rust front door take the entry point while unported units fall back to the original, then lift units until replay is 100% in pure Rust. The README claims 38,251 recorded cases replayed with zero mismatches across its demos and speedups from 3x (Java slugify) to 20x (a TypeScript search over 50,000 keys).
Agent workflow around it
The whole-app skill structures multi-session agent work with files: decisions as ADRs in rust/DECISIONS.md (settled by an up-front question round with a recommended answer for each), a glossary mapping original terms to Rust names, background research notes per dependency, a rust/MAP.md of every unit and its state that a session claims before starting, a rust/HANDOFF.md at each checkpoint, and a generated cutover.sh for steps only a person can do (deploys, DNS, secrets), each with a way back.
What it can't catch (stated by the project)
Inputs nobody recorded: the recording is the ceiling, so add production samples, bug reports and the original's fixtures.
Regex semantics: the Rust
regexcrate lacks lookaround and backreferences, and Unicode classes and case folding differ; record adds traps, but real patterns need fuzzing on real text.Nondeterminism: values that change between runs become wildcards and reordered lines compare in any order, which can hide a real ordering or concurrency change.
Float formatting differences between languages; tolerance forgives rounding, never formatting.
Anything outside output: timing, memory, concurrency under load, network calls, databases and signals are not compared.
Tradeoffs
Recorded-behavior equivalence is a strong, cheap oracle but only as good as its inputs; it encodes the original's bugs as the spec.
The bridge phase means two runtimes in one process for a while, with FFI overhead and two sets of failure modes.
Very young project: a day old at the time of the post, with broad language claims that are not yet battle-tested.
Review checklist
Do the recorded inputs include production samples and past bug reports, not only generated cases?
Were wildcarded and order-insensitive diffs reviewed one by one?
Are side effects (files, network, database writes) covered by an app-level replay or separate tests?
Did bench show a real win on a realistic workload before porting, and after batching?
Is shadow mode run long enough to see rare inputs before the gate is trusted?
Is the rollback path to the original kept until cutover is complete?
Related entries
R082 Craft Anything: rebuild software, graded against the original.
R080 AWS Emulator: a replica graded by the real service.
R020 The model proposes, the harness enforces.