Code language detector
Paste a source snippet and get the likely language, confidence, runner-up candidates and the signals that led to the result. Runs locally in your browser.
About this tool
Paste a code snippet to identify its likely programming language without uploading the text or calling a model. The detector uses a deterministic, GitHub-Linguist-style score table: syntax signals, imports, keywords, shebangs, structural checks, and optional filename hints add weighted evidence for each language. The result includes a confidence level, ranked alternatives, and the matched signals so you can see why a language won.
Worked example: paste this Rust snippet and set the filename hint to main.rs:
use std::collections::HashMap;
pub fn tally(words: &[&str]) -> HashMap<String, usize> {
let mut counts = HashMap::new();
for w in words {
*counts.entry(w.to_string()).or_insert(0) += 1;
}
counts
}
The report should identify rust, show the filename and Rust-specific signals in the evidence,
and list the nearest alternatives. For a short one-liner such as print("hello"), use the candidate
allowlist (python,javascript,ruby) or a filename hint to make the ambiguity explicit.
Limits and edge cases: snippets are capped at 1 MiB, the detector is heuristic rather than a trained model, and very short or deliberately polyglot snippets can be close calls. Syntax highlighting and file upload are intentionally separate from this tool; this page focuses on local paste-in language detection with explainable output.
FAQ
How accurate is the detector?
It is strongest on snippets with several lines, imports, type declarations, markup structure, or a filename hint. It is not an ML classifier, so it does not claim corpus-level accuracy; instead it shows the signals that matched and warns when the top candidates are close.
Why does a one-line snippet get a low-confidence warning?
Many languages share tiny snippets. print("hello") could be Python, Ruby-like pseudocode, or a
function call in another language. Add more context, set filename, enable common_only, or provide
a candidates allowlist to make a short snippet less ambiguous.
What does the candidate allowlist accept?
Use comma-separated language ids such as rust,python,javascript or json,yaml,toml. Unknown ids
produce an error that lists the supported values, and the report notes when the language pool was
restricted.
Does this upload my code or use a network service?
No. The scoring runs in local WebAssembly in the browser and the CLI/chat block uses the same pure Rust core. The pasted snippet is not sent to a remote detector.
Developer & Automation Access
Run it from the terminal
Same engine as this page, headless — via the gizza CLI:
gizza tool code-language-detect 'Paste code here, for example:
fn main() {
let mut count = 0;
println!("{count}");
}'New to the CLI? Get gizza →
Open it by URL
Pre-fill and auto-run this tool with query parameters — the names match the API/CLI:
https://gizza.ai/tools/code-language-detect/?code=Paste%20code%20here%2C%20for%20example%3A%0Afn%20main%28%29%20%7B%0A%20%20%20%20let%20mut%20count%20%3D%200%3B%0A%20%20%20%20println%21%28%22%7Bcount%7D%22%29%3B%0A%7D&filename=main.rs&candidates=rust%2Cpython%2Cjavascript&top_k=3&common_only=true&explain=true&output=reportMachine-readable descriptor: tool.json — title + parameters JSON Schema for agents.