A MOSS alternative that also checks the web and AI
MOSS (Measure of Software Similarity) compares the files in one batch with each other, and it does that well. It does not search GitHub or Stack Overflow, has no AI detection, and deletes its results after 14 days. Codequiry runs the same comparison and adds the rest.
2-day free trial. Checks every file against classmates, the web and AI, in 65+ languages.
Lab3.javaanswerA distinctive one-line trick, copied with the names changed (highlighted). Peer comparison cannot see it, because the original was never in the batch.
Used in programming courses at
What MOSS does well
MOSS was written by Alex Aiken at Stanford in 1994 and it is still the tool many CS departments start with. It earned that place:
- It is free for academic, non-commercial use.
- It ignores cosmetic edits. Its winnowing fingerprints skip names, comments and whitespace, so a renamed copy of a classmate's program still matches.
- It handles starter code. Pass the base file and code every student was given is left out.
- It is fast on a batch of a few hundred files.
If the only question is "did two students in this class hand in the same program?", MOSS answers it well.
Where MOSS falls short
Most copying in a programming course no longer happens between classmates. It comes from the web and from AI, and MOSS was never built to look there.
- No web search. A solution copied from GitHub or Stack Overflow passes unless another student copied the same one. The example above is exactly that case.
- No AI detection. Code from ChatGPT, Claude, Gemini or Copilot passes unless two students handed in nearly identical output.
- Results expire. The report page is deleted after 14 days. If a case reaches a hearing later in the term, you need your own copy.
- A script, not an app. You register by email, run a Perl script from a terminal, and read results as linked HTML pages. There is no dashboard, no LMS integration and no history across terms.
- No support. It is a free research service. When the server is slow or down, there is no one to call.
Peer comparison has limits inside the batch too. What MOSS misses shows refactored copies it does not flag, side by side.
MOSS vs Codequiry
| Stanford MOSS | Codequiry | |
|---|---|---|
| Compares students with each other | Yes | Yes |
| Excludes starter code | Yes | Yes |
| Checks GitHub, Stack Overflow and the web | No | Yes |
| Detects AI-generated code | No | Yes |
| Compares with past terms | No | Yes |
| Languages | About 25 | 65+ |
| How you use it | Perl script | Web app, LMS, API, CLI |
| Results kept | 14 days | In your account |
| Cost | Free for academic use | Paid plans, free trial |
Comparing more tools? MOSS vs Dolos vs JPlag covers the other free options.
Switching from MOSS
You do not have to change how you collect submissions. The same folder of files you send to MOSS works here.
- Create an assignment and pick the language. Add the starter code if there is any.
- Upload the submissions as a zip, one folder or file per student, or pull them from your LMS.
- Choose the checks. Peer comparison runs on every check. Add web sources and AI detection, or pick our independent winnowing engine for fingerprint-based peer comparison. Scores from different tools are not interchangeable.
- Read the results. Each student gets a score, and each match opens the two files side by side with the shared lines marked.
Run your last MOSS batch through Codequiry
Upload the same files and see what the web and AI checks add.
2-day free trial
When MOSS is still enough
MOSS is a fine choice for a small class, an assignment unique enough that no answer exists online, a course that allows AI, and an instructor comfortable in a terminal. If all four are true, keep using it.
If any of them is not, peer comparison alone will miss copies. That is the gap a code plagiarism checker with web and AI detection closes. For a one-off check without an account, the free checker compares files with each other at no cost.
Frequently asked questions
Is MOSS free?
Yes, for academic, non-commercial use. You register by email with Stanford, receive a submission script, and send files from the command line. Commercial use is not covered.
Does MOSS check GitHub or Stack Overflow?
No. MOSS only compares the files you send it with each other. A solution copied from the web is caught only if another student in the same batch copied the same solution.
Does MOSS detect AI-generated code?
No. MOSS has no AI classifier. Code written by ChatGPT, Claude, Gemini or Copilot passes unless two students handed in nearly the same generated answer.
How long do MOSS results last?
MOSS typically keeps results for 14 days, but may delete them earlier. Save the report if it is needed for a review.
Can Codequiry run a MOSS-style comparison?
Yes. Codequiry uses independently developed winnowing-based peer comparison alongside Hexagram and Zeus. It is separate from Stanford MOSS, and can add web and AI checks to the same report.
Can I exclude starter code, like the MOSS -b option?
Yes. Upload the starter or base code with the assignment and matches against it are left out, the same idea as MOSS's base file option.
Which languages are supported?
MOSS supports about 25 languages. Codequiry supports 65+, including Python, Java, C, C++, C#, JavaScript, TypeScript, Go, Rust, Kotlin, Swift, SQL and Jupyter notebooks.
Keep the peer check. Add the web and AI.
Upload a class's submissions and see which ones match classmates, the web or AI, with the lines side by side.
Start free trial2-day free trial
Just one assignment? Run a single check