Any other future supported language?

[educhet-inovertis]

This tool seems to be a very good idea and concept in our age of programming, I sadly don't use TS or JS myself and I was wondering if there was any plan to add other languages or structure (C++ for example) to the project or any similar project out there?

[BartWaardenburg]

Thanks for the kind words, and good question.

No plans for other languages, I'm afraid. Fallow is built on the Oxc parser ecosystem, which is JS/TS-specific, and the analysis is shaped around JavaScript/TypeScript semantics: ESM/CommonJS imports, re-export chains, package.json dependencies, tsconfig path aliases, and the framework plugins. Almost none of that maps onto a language like C++, so it would really be a separate tool rather than a feature here.

There are good options elsewhere, though. The ones closest to what fallow does (covering several concerns at once across many languages) are:

• SonarQube, which spans 30+ languages including C++, with duplication, complexity, code smells, and dependency checks
• CodeScene, which does behavioral hotspot analysis (complexity weighted by how often code changes) plus coupling and architecture views

If you'd rather mix focused tools, by concern:

• Duplication: PMD CPD and jscpd both handle C++ and many other languages
• Complexity: lizard computes cyclomatic complexity across most languages
• Architecture / circular deps: NDepend (.NET), ArchUnit (JVM), pydeps (Python)
• Unused includes / code: for C++, include-what-you-use and cppcheck

For C++ specifically, SonarQube or CodeScene will get you the broad picture, with lizard, PMD CPD, IWYU, and cppcheck filling in the individual pieces.

Worth adding, since SonarQube and CodeScene are the closest comparison for fallow itself, where it differs on a JS/TS codebase. Fallow is deterministic codebase intelligence: it reports a quality score, PR risk, complexity hotspots, duplication, architecture and boundary issues, dependency hygiene, and cleanup opportunities, all derived from one module graph in a single pass. A few things set it apart:

• Deterministic, with no LLM in the analyzer. Same input, same output, every run, which makes it predictable in CI and trustable for AI agents to act on. Agents call fallow for facts rather than fallow guessing. As more reviewers lean on LLMs, that matters more, not less.
• One graph, broad integration. The same graph feeds cycles, duplication, boundaries, hotspots, and dependency hygiene, so findings are connected rather than siloed across separate rule sets. It's the broadest-integrated module graph in open-source JS/TS tooling.
• Rust-native and local. A single binary that runs offline in sub-second time with zero config: no server, account, or hosted backend to stand up. It's free and open source.

All three now ship MCP servers for AI agents, so that part isn't unique. The difference is that fallow's just wraps the local binary, with nothing else to provision.

If you ever do touch JS/TS, would love to hear how fallow works out for you.