AI Code Detector Runs in your browser

AI CODE DETECTOR

AI Code Detector for transparent code analysis.

Paste code, upload source files, or choose a local folder to review configured code-style signals and their flagged lines.

Detection runs in your browser. Your code is not uploaded.

No files selected. Up to 100 supported files, 1 MB each, 5 MB total.

What this tool is

AI Code Detector is a browser-based code-style review. Paste source into the editor or choose a local source file; the review runs in the current browser and this static site does not upload your code to a server.

How this heuristic works

The published configuration checks six code-style signals: explicit comments that name AI tools, boilerplate explanatory comments, long documentation-style comments, highly descriptive multi-word identifiers, repeated adjacent comments, and template-style defensive error text. Each matched rule contributes its configured weight to the displayed heuristic score.

Code style is a clue, not evidence. A matching signal cannot prove who wrote the code, establish that an AI tool was involved, or settle an authorship question. Read it as a prompt to inspect the surrounding code, not as a verdict.

Supported input and languages

You can paste code, select local source files, or choose a local folder. The language picker supports Python, JavaScript, Java, C++, PHP, and C#. The file picker accepts Python files; JavaScript files including JSX and module files; Java files; common C++ file extensions; PHP files; C# files; and plain-text files. To keep the local review responsive, choose up to 100 supported files, no more than 1 MB per file and 5 MB in total; each source file can contain up to 10,000 lines. When several supported files are selected, each file receives its own local heuristic result.

How to read the report

The score is the sum of the configured weights for rules that matched your submitted code. It describes the presence of those configured signals, not a conclusion about the source of the code.

Matched signals name the rules that fired and point to the relevant lines. Highlighted lines show every line involved in a match. Use both to review the local context: a comment can be descriptive for a valid reason, and an identifier can be clear because a team chose a naming convention.

After an analysis, you can download the displayed local results as a JSON report. The download is generated in this browser.

When to question the result

Do not rely on this score when style overlap is the main explanation. People can learn from one another, teams can share templates, and linters, formatters, generators, and project conventions can produce similar patterns. Check the surrounding implementation, project history, and the reason for each matched signal before making a judgment.

Related ways to find this page

What is an AI code checker? It is another way to describe this local review for matched code-style signals and highlighted lines.

What is a code AI detector? Here it means a transparent browser-based heuristic for inspecting style signals, not a tool that determines authorship.

What is an AI generated code detector? This page checks configured style signals and reports the lines that caused each signal to match.

How to tell if code is ai generated? Review the local report together with code history and surrounding implementation; this heuristic is not proof of authorship.

LEARN

Read the report with the right context

These short guides explain what a matched rule means and where this local heuristic stops.

MORE GUIDES

Use a focused guide when the question is about the scope of a local code-style review rather than an authorship verdict.

About this local code reviewSee the tool’s stated scope and limits. Read the privacy policySee what happens when you inspect code in the browser. Contact the developerFind the site contact details and useful context to include. Review boilerplate and AI-code claimsUnderstand why a repeated pattern is not a source claim. Check claims about ChatGPT or CopilotLearn why style alone cannot name a code source. Separate AI signals from code safetySee why security requires a different review. Review a partial code snippetKeep a focused input in its proper scope. Use the detector in pull-request reviewTurn a matched rule into a concrete code question. See supported programming languagesCheck the languages available in the local picker. Understand flagged commentsRead the specific rule before interpreting a comment. Understand why a score changesCompare the changed input and matched rules. See what the JSON report containsCheck which result details are exported and what is not included. Check whether a score is repeatableCompare identical inputs under the same configured rules. Understand flagged line numbersSee how report line references map to submitted code. Check screenshot input limitsSee why the detector needs code text rather than an image. Check LF and CRLF inputSee how common newline formats map to the same line-based checks. Check Unicode identifier signalsSee the current rule's limits with non-ASCII names. Check Markdown code fencesSee how pasted fence lines affect the scanned input and line references. Check indentation and scoresSee what the current examples show when leading spaces change. Check signals inside string literalsSee why a text pattern can match even when it is not a comment.

BROWSER TOOL

Convert a visual asset locally

Working with an SVG image alongside your code? Open the separate converter below to make a PNG in your browser.

Your report will appear here

We only inspect the code in this browser tab. Nothing is uploaded by this demo.