Live detection feed

Your students are copying code.
We catch all of it.

MOSS rebuilt with winnowing and AI detection. Upload once — catch peer copying, ChatGPT code, StackOverflow pastes, and Chegg answers in the same check.

Start free trial See it in action
student_47.py flagged 92% match
Peer match detected with student_12.py — BFS implementation identical with renamed variables
CS188 Pacman Project • 2 min ago
AI-generated code 78% GPT
Hexagram flagged submission — instructional comments, over-documented, textbook-perfect DP pattern
CS201 HW3 • 5 min ago
Web source found 96% SO
Merge sort implementation copied verbatim from stackoverflow.com/questions/6567941
CS101 Sorting • 8 min ago
Chegg answer detected 87%
BST construction matches homework answer on chegg.com/homework-help — 3 students submitted same code
CS50 Problem Set 5 • 12 min ago
20+
Detection engines
1T+
Sources indexed
65+
Languages
500+
Universities

See exactly what gets caught

Real examples from CS courses. Click the tabs to explore each detection type.

Peer Match 92%
AI Detection 78%
Web Source 96%
Network View cluster
student_a / search.pySource
14def breadthFirstSearch(problem):
15 """Search the shallowest nodes."""
16 frontier = util.Queue()
17 frontier.push((problem.getStartState(),
18 [], 0))
19 explored = set()
20 while not frontier.isEmpty():
21 node, actions, cost = frontier.pop()
22 if problem.isGoalState(node):
23 return actions
24 if node not in explored:
25 explored.add(node)
student_b / search.pyMatch
11def bfs(prob):
12 """BFS search."""
13 q = util.Queue()
14 q.push((prob.getStartState(),
15 [], 0))
16 visited = set()
17 while not q.isEmpty():
18 state, acts, c = q.pop()
19 if prob.isGoalState(state):
20 return acts
21 if state not in visited:
22 visited.add(state)
92% structural matchVariable renamingSame algorithm
CS188 Pacman Project — BFS
submission.pyAI flagged
1# Elegant and efficient DP solution
2def longest_common_subsequence(text1, text2):
3 """Find the LCS using dynamic
4 programming for optimal performance."""
5 m, n = len(text1), len(text2)
6 dp = [[0]*(n+1) for _ in range(m+1)]
7 for i in range(1, m + 1):
8 for j in range(1, n + 1):
9 if text1[i-1] == text2[j-1]:
10 dp[i][j] = dp[i-1][j-1]+1
11 else:
12 dp[i][j] = max(dp[i-1][j],
Hexagram Analysis78% GPT-4
Detection Signals
Instructional commentLine 1
Over-documented docstringLines 3-4
Uniform variable namingLines 5-6
Textbook-perfect structureLines 9-12
Verdict: AI-generated
ChatGPT 78% • Claude 12% • Gemini 4%
78% ChatGPT12% ClaudeHexagram v2
CS201 — Dynamic Programming
student_submission.jsWeb match
1function mergeSort(arr) {
2 if (arr.length <= 1) return arr;
3 const mid = Math.floor(arr.length / 2);
4 const left = arr.slice(0, mid);
5 const right = arr.slice(mid);
6 return merge(mergeSort(left),
7 mergeSort(right));
8}
9
10function merge(left, right) {
11 let result = [], i = 0, j = 0;
12 while (i<left.length && j<right.length){
Matched Sources4 found
stackoverflow.com
/questions/6567941/merge-sort-js
96%
Lines 1-7
github.com
trekhleb/javascript-algorithms
44%
Lines 10-12
chegg.com
/homework-help/sorting-impl
31%
Lines 1-5
coursehero.com
/file/cs101-sorting-assignment
87%
Lines 1-7
96% StackOverflow44% GitHub87% CourseHero
CS101 — Sorting Algorithms
A B C D E F G 92% 41% 87%
Cluster detected: Students A, B, D
3 submissions share 87-92% structural similarity. Likely collusion ring.
High risk Medium Low Clean
Collusion cluster3 students flagged4 clean
CS188 — 47 submissions analyzed

20+ engines run on every single check

MOSS is one of many. Every submission is analyzed by all engines simultaneously.

M

MOSS

Stanford's Measure of Software Similarity. Rebuilt with winnowing fingerprinting and multi-threaded processing on dedicated infrastructure.

Unlimited checks25+ languages
H

Hexagram

Proprietary AI detection engine. Identifies code written by ChatGPT, Claude, Gemini, and Copilot through pattern analysis of commenting style, structure, and naming.

v2.0.1AI models
Z

Zeus

Structural code comparison engine. Compares every submission against every other in the class. Catches variable renaming, reordering, and obfuscation attempts.

Peer-to-peerCluster detection
W

Web Search

Scans 1T+ indexed sources including StackOverflow, GitHub, Chegg, CourseHero, Brainly, GitLab, Bitbucket, and academic repositories.

1T+ sourcesReal-time
D

Dolos

Token-based analysis engine. Fast, language-agnostic comparison that catches structural similarities even when surface-level code looks different.

65+ languagesFast
J

JPlag

Graph-based comparison with token sequence matching. Industry-standard engine used by universities worldwide alongside Codequiry's own engines.

Graph analysisEstablished

Built for how professors actually work

Not just detection — a complete workflow from upload to academic hearing.

Side-by-side comparison

Line-by-line highlighted matches

Every match shows the exact lines, highlighted and aligned. Click any match to see the source — whether it's another student, a web page, or an AI pattern.

student_a.py
pivot = arr[mid]
left = [i for i in arr
if i < pivot]
student_b.py
p = lst[middle]
lo = [x for x in lst
if x < p]
Risk scoring

Estimated unoriginality

Every submission gets a composite risk score combining peer match, web match, and AI detection into one number.

84
HIGH RISK
Peer 92%Web 64%AI 78%
Integrations

LMS + API + CLI

Connect Canvas, Moodle, or Blackboard via LTI. Or use the REST API with SDKs for Node.js, Python, Java, Go, and PHP.

Canvas Moodle Blackboard REST API CLI
Reports

PDF reports for academic hearings

Generate detailed, printable PDF reports with all evidence — matched lines, source URLs, AI confidence scores, and similarity percentages. Ready for academic integrity committees.

PDF Export CSV Export Print
Team

50,000 seats

Admin, TA, and instructor roles. Bulk import students, share results across your department.

Storage

100GB included

File manager with folders, starring, trash, and recovery. All results stored permanently.

Compliance

FERPA + GDPR

SOC 2 Type II certified. Full audit logs, data disclosure tracking, and student privacy controls.

Stanford MOSS vs Codequiry

Same algorithm at its core. Everything else is better.

CapabilityStanford MOSSCodequiry
SubmissionsRate limitedUnlimited
AI detectionNoneChatGPT, Claude, Gemini
Web searchNone1T+ sources
ProcessingSingle-threadedMulti-threaded, dedicated
Cluster detectionNoneNetwork graph analysis
InterfacePerl scriptWeb, API, CLI, Desktop, LMS
Results14 daysPermanent + PDF export
Languages~2565+
LicenseAcademic onlyCommercial

Stop chasing MOSS limitations.
Start catching everything.

Included with the Advanced plan. 7-day free trial, no charge until it ends.

Start free trial Advanced plan • $200/mo after trial • MOSS + AI + Web + 20 engines