Three scans. One platform.
Unlimited peer comparison.
MOSS Winnow is Codequiry's large-batch engine, built on our own tokenization system. All-in-one source code plagiarism detection for peer copying, web copying, and AI-generated code detection. Group similarity, web plagiarism, and AI detection in a single check with native multi-threading.
The Power of Three
Every check runs three distinct scans simultaneously, detecting peer copying, web plagiarism, and AI-generated code in a single operation.
Winnowing Group Similarity
Compare all student submissions against each other to detect peer-to-peer code copying with precision matching.
- All-pairs comparison
- Variable/structure renaming detection
- Winnowing fingerprint algorithm
- Similarity percentage per pair
Web Plagiarism for Code
Scan every submission against online sources, including GitHub, Stack Overflow, code repositories, tutorials, and more.
- indexed code sources
- GitHub & Stack Overflow matching
- Real-time web crawling
- Source URL attribution
AI Detection for Code
Identify AI-generated code from ChatGPT, Claude, Copilot, and other AI assistants using Codequiry's own AI detection analysis.
- 90% AI/GPT detection, 1.3% false flags
- GPT-4, Claude, Copilot detection
- Line-level AI probability
- Multi-model coverage
Languages Supported
Python, Java, C/C++, JavaScript, Go, Rust, Ruby, PHP, Swift, Kotlin, and 55+ more.
Unlimited Usage
No per-check fees. Unlimited group similarity, unlimited assignments, unlimited submissions.
Native Multi-Threading
Process all submissions in parallel with native threading. Optimized for parallel processing on Codequiry infrastructure.
How It Works
Four simple steps from upload to results. Multi-threaded processing handles everything in parallel.
Upload Submissions
Upload student code files as a ZIP or individually. Supports 80+ submissions per check.
Select Scans
Choose group similarity, web plagiarism, AI detection, or run all three simultaneously.
Multi-Threaded Processing
Native multi-threading processes every submission in parallel for maximum speed.
Review Results
View detailed reports with similarity scores, matched sources, line-by-line comparisons, and AI probability.
Built for Speed
Native multi-threading and local processing deliver results faster than any remote MOSS service.
Process 80+ submissions in under 2 minutes
- Native multi-threading, every submission processed in parallel threads, not sequentially
- Parallel processing on Codequiry infrastructure
- Native performance, an optimized engine runs on Codequiry's infrastructure
- Under 2 minutes for a full 3-scan check with 80+ submissions
- Dedicated winnowing server, on Codequiry's own infrastructure
How MOSS Winnow Compares
See how MOSS Winnow stacks up against Stanford MOSS and free MOSS online tools.
| Capability | MOSS Winnow | Stanford MOSS | Free MOSS Online |
|---|---|---|---|
| Group Similarity | Yes | Yes | Limited |
| Web Plagiarism Scan | web sources | No | No |
| AI Detection for Code | 90% AI/GPT detection | No | No |
| Submission handling | 80+ | 100 submissions/day/user | Varies |
| Multi-Threading | Native | Not specified | No |
| Unlimited Usage | Unlimited | Rate limited | No |
| Languages | 65+ | ~25 | ~10-15 |
| Results Format | Web UI + API + JSON | HTML only | HTML only |
| Setup Required | None, instant | Perl + email registration | Varies |
Who It's For
MOSS Winnow serves educators and institutions who need reliable, scalable code plagiarism detection.
CS Departments
Run MOSS Winnow across all programming courses with unlimited checks, 65+ languages, and three-in-one scanning.
Teaching Assistants
Grade faster with instant plagiarism reports. Upload submissions, run checks, and review results in minutes.
Programming Contests
Verify submission originality in competitive programming events with fast, accurate similarity detection.
Bootcamps & MOOCs
Scale code integrity checks across hundreds of online students with unlimited usage and API access.
Academic Integrity Offices
Investigate suspected plagiarism cases with detailed similarity reports, source attribution, and AI detection.
Research Labs
Ensure research code originality and detect unauthorized code reuse across research group submissions.
One Purchase. Unlimited peer comparison.
Pick the plan that fits your needs. See the current plan details for engine availability.
For teams getting started with code integrity.
- Checks & Detection
- 50 web + AI scans/month · ∞ submissions
- Unlimited Group Similarity
- AI-Written Code Detection
- Scan GitHub, StackOverflow, Chegg & the indexed web
- Engines
- Hexagram™ v2.0.1 Fast
- Dolos + JPlag included
- 65+ programming languages
- Platform
- Web Dashboard + Desktop Client
- API/CLI access
Better detection engines, more scanning power, more web sources.
- Checks & Detection
- 75 web + AI scans/month · ∞ submissions
- Unlimited Group Similarity
- AI-Written Code Detection
- Scan GitHub, StackOverflow, Chegg & the indexed web
- Engines
- Hexagram™ v2.0.1 Fast
- MOSS Limited
- Dolos + JPlag included
- Custom engine builder
- 65+ programming languages
- Scale & Limits
- Unlimited student seats & submissions
- Unlimited assignments
- No upload limits
- Platform
- Web Dashboard + Desktop Client
- Unlimited API/CLI access
- LMS integrations
- 2 admin seats
Frequently Asked Questions
MOSS Winnow is Codequiry's large-batch engine, built on our own tokenization system, which took inspiration from winnowing and adds stronger false-positive reduction, more languages, bulk processing and better base-code detection. It includes three scans in one, winnowing-based group similarity, web plagiarism detection, and AI-generated code detection, with native multi-threading and support for 65+ programming languages. Peer comparison, web checks, and AI analysis follow the selected plan.
Stanford MOSS provides submission similarity checking, with a public-service limit of 100 submissions per day per user. MOSS Winnow adds web plagiarism scanning across GitHub, Stack Overflow, and the indexed web, AI-generated code detection, supports 80+ simultaneous submissions, uses native multi-threading, and provides peer comparison under your plan; web and AI checks use credits.
The three scans are: (1) Winnowing Group Similarity, which compares all submissions against each other to detect peer-to-peer copying, (2) Web Plagiarism for Code, which scans against online sources including GitHub and Stack Overflow, (3) AI Detection for Code, which identifies AI-generated code from GPT, Claude, Copilot, and other AI models.
MOSS Winnow supports 80+ simultaneous submissions per check. With native multi-threading, all submissions are processed in parallel for maximum speed, so most checks complete in under 2 minutes.
Yes. The Advanced plan includes unlimited group similarity checks, unlimited assignments, and unlimited submissions. Web and AI scans use included credits, with the option to add the Rapid add-on for 10 web + AI scans/mo on dedicated infrastructure.
MOSS Winnow supports 65+ programming languages including Python, Java, C, C++, JavaScript, TypeScript, Go, Rust, Ruby, PHP, Swift, Kotlin, R, MATLAB, SQL, HTML/CSS, and many more.
MOSS Winnow uses native multi-threading on Codequiry infrastructure. Runtime depends on file count, code size, and the analysis selected.
The AI detection uses Codequiry's v3e model, trained on Claude Opus 5.5, GPT-5.6 Sol and GPT-5.5 Codex submissions. It catches 90% of AI/GPT-written code and falsely flags 1.3% of human-written code, and flags code from ChatGPT, Claude, GitHub Copilot, Gemini and other AI coding assistants.
No. MOSS Winnow runs on Codequiry's own engines and servers.
MOSS Winnow is included in the Codequiry Advanced plan. Visit the pricing page, select the Advanced plan, and you'll have immediate access to all three scans, unlimited peer comparison, and the full Codequiry platform.
Stop guessing. Start detecting.
Three scans in one check. Unlimited peer comparison. Native multi-threading. The most complete code plagiarism detection available.
What instructors say
-
Codequiry protects our course from copying. It's the only reliable tool we've found.
Course instructor -
Codequiry has been the best experience I've had checking source code against the web for plagiarism. It even covers Stack Overflow!
Codequiry user -
As a professor at Stanford, I want to make sure students submit code they wrote themselves. Codequiry checks their code against the web for plagiarism, something other checkers can't do.
Professor, Stanford University
Tools referenced on this page
Codequiry is compared with these tools above. Each card links to the project's official site, so you can check the claims yourself.
- MOSS Stanford University, Alex Aiken Free academic similarity service from 1994 that fingerprints code with winnowing and reports matching passages between submissions. theory.stanford.edu (opens in a new tab)
- JPlag Karlsruhe Institute of Technology, open source Open-source token-based detector that compares programs with Greedy String Tiling and produces a similarity report per pair. github.com (opens in a new tab)
- Dolos Ghent University, open source Open-source similarity tool built on tree-sitter parsing, with a web interface for exploring clusters of related submissions. dolos.ugent.be (opens in a new tab)