Codequiry AI code detector
Vendor CQC Pvt. Ltd. (Codequiry) Official site codequiry.com/detect-ai-written-code How this card is sourced
Updated · AI Detector Index editors
Codequiry states that its current model, v3e, catches 90% of AI-written code and falsely flags 1.3% of human-written code, on the Codequiry AI code detection page read 9 November 2026. The vendor is CQC Pvt. Ltd. of Chennai, India, founded in 2017, and the product is the first card in the detector index whose only modality is source code. How this card is sourced applies the same fourteen fields to code that the other cards apply to text.
What Codequiry detects
| Field | Value | Source |
|---|---|---|
| Content types detected | Code | codequiry.com/detect-ai-written-code |
| Generators covered | ChatGPT (GPT-5.5, GPT-5.5 Instant, GPT-5.4 Pro), GitHub Copilot, Claude (Opus 4.8, Sonnet 4.6, Haiku), Grok (Grok 4.3, Grok Code), Google Gemini (3.5 Pro, 3.5 Flash), Cursor AI, Amazon Q Developer (CodeWhisperer), Meta Llama (Llama 4 Maverick, Scout), General AI Detection model for any AI-generated code patterns | codequiry.com/detect-ai-written-code |
| Input limits | Source code by drag and drop, paste, ZIP upload or LMS import, in 65+ programming languages (Python, Java, JavaScript, C++, C#, PHP, Ruby, Go, Rust, Swift, SQL, R and 38+ more). Checks are metered per submission: Grow includes 20 web + AI checks per month, then C$6.80 per check as displayed. AI detection takes 10 to 30 seconds per submission. No file size limit published. | codequiry.com/detect-ai-written-code |
| Free tier | Codequiry Free: $0/mo, unlimited peer similarity checks within a class, no card needed; web, GitHub and AI detection are not included, so AI code detection has no free tier | codequiry.com/pricing |
| API | Yes, codequiry.com/usage/api | codequiry.com/usage/api |
| Integrations | Canvas (bulk import), Moodle (bulk import), Blackboard (bulk import), LTI LMS integration, Desktop client, Command-line client (Plagiarism Terminal), MCP server (Claude, Cursor), SSO/SAML, ZIP upload | codequiry.com/detect-ai-written-code |
| Data retention | Site footer and privacy policy (last updated October 22, 2026, version 5.0): source code submitted on any plan without an active Rapid add-on is added in anonymised form to a cross-account matching corpus, is retained indefinitely including after account closure, and cannot be deleted on request; content submitted while Rapid is active is excluded from the shared corpus and can be deleted on demand. The AI detector page FAQ says: You maintain full control over your data and can delete submissions at any time. | codequiry.com/privacy-policy |
| Launch year | 2017 | codequiry.com/about |
| Ownership | CQC Pvt. Ltd., Infra Limited, 200 Feet Radial Rd, MCN Nagar Extension, Pallavaram, Thoraipakkam, Chennai, Tamil Nadu 600097, India; founded 2017 (about page). A San Francisco office address is listed as not accepting correspondence. The service is not offered in the EU, EEA or UK. | codequiry.com/privacy-policy |
Codequiry detects AI-generated source code and code plagiarism, and the two run in one report. The detection page names ChatGPT (GPT-5.5, GPT-5.5 Instant and GPT-5.4 Pro), GitHub Copilot, Claude (Opus 4.8, Sonnet 4.6 and Haiku), Grok, Google Gemini, Cursor AI, Amazon Q Developer and Meta Llama, each with a stated detection rate between 80% and 91%, plus a "General AI Detection" model for any AI-generated code pattern (codequiry.com/detect-ai-written-code, read 9 November 2026). Coverage is stated as 65+ programming languages, with Python, Java, JavaScript, C++, C#, PHP, Ruby, Go, Rust, Swift, SQL and R named and "38+ more" including MATLAB, Kotlin, TypeScript, Scala and Perl.
The languages field on every card counts natural languages, so Codequiry's 65 programming languages sit in the input limits field instead, and the card carries no natural-language count. Integrations named on the page are bulk import from Canvas, Moodle and Blackboard, an API, SSO/SAML and ZIP upload; the about page adds LTI, a command-line client, a desktop client and an MCP server for Claude and Cursor (about page, read 9 November 2026). The API is documented at codequiry.com/usage/api.
The company is CQC Pvt. Ltd., with its corporate headquarters at Infra Limited, Chennai, Tamil Nadu, and a San Francisco address that the footer says accepts no correspondence (privacy policy, read 9 November 2026). The site footer states that the service "is not offered in, and may not be accessed or used from, the European Union, the European Economic Area or the United Kingdom". The two text-first cards that also list code are Copyleaks and Originality.ai.
Limits and plans
| Plan | Price | Limits | Checked on |
|---|---|---|---|
| Grow | C$272 /mo as displayed to a reader in Canada (the page's currency script converts a US dollar base at 1.36, which makes Grow US$200; the AI detector page states plans from $200/month) | 20 web + AI checks per month, C$6.80 per check after; AI-written code detection; Hexagram similarity engine; unlimited peer checks; 20 HyperScans handed over once; 1 instructor seat; web dashboard, desktop client and API/CLI. | |
| Advanced (add-on to Grow) | C$406.64 /mo add-on as displayed (US$299 base) | Adds 20 web + AI checks per month; 2 instructor seats; LMS integrations; winnowing detection; 40 HyperScans. | |
| Defense (add-on to Advanced) | C$950.64 /mo add-on as displayed (US$699 base) | Adds 125 web + AI checks per month; Hexagram and Zeus plus winnowing engines unlimited; dedicated server; 20 instructor seats; custom LMS and SIS integrations; audited hosting, DPA and custom MSA; commercial use allowed; 125 HyperScans. | |
| Rapid (add-on) | C$40.80 /mo add-on as displayed (US$30 base) | Adds 10 web + AI checks and 10 HyperScans per month; priority queue; the right to delete assignments and submissions, which never enter the matching database. | |
| Enterprise | Contact sales; enterprise pricing available. |
Checks are metered per submission rather than per word. The Grow plan includes 20 web and AI checks a month and charges C$6.80 per further check as displayed; AI detection takes 10 to 30 seconds per submission; no file size limit is published (codequiry.com/pricing and the detection page, read 9 November 2026). The free plan, Codequiry Free at $0 a month, covers unlimited peer-to-peer similarity checks within a class and excludes web, GitHub and AI detection, so AI code detection has no free tier.
The pricing page displayed Canadian dollars to this reader; the page's own currency script converts a US dollar base at a rate of 1.36, and the detection page states "Plans from $200/month". Figures below are as displayed, read 9 November 2026.
Retention is the field to read twice. The footer and the privacy policy of October 22, 2026 state that code submitted on any plan without an active Rapid add-on "is added in anonymised form to a cross-account matching corpus, is compared against future submissions, is retained indefinitely, including after account closure, and cannot be deleted on request"; only Rapid customers can delete submissions (privacy policy, read 9 November 2026). The detection page FAQ says the opposite in fewer words: "You maintain full control over your data and can delete submissions at any time." The card carries both sentences.
What the vendor claims
Our current model, v3e, is 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.
Codequiry's accuracy statement reads: "Our current model, v3e, is 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" (codequiry.com/detect-ai-written-code, read 9 November 2026). The per-tool rates on the same page are ChatGPT 90%, GitHub Copilot 91%, Claude 84%, Grok 83%, Google Gemini 88%, Cursor AI 81%, Amazon Q 80% and Meta Llama 80%.
On false positives the vendor says: "Human-written code rarely trips it: 1.3% is falsely flagged. That is low, not zero, so every flag shows the lines it rests on and you can check them before acting" (same page, read 9 November 2026). No sample size accompanies either figure, and the footer adds that "statistics, counts and performance figures are approximate, may be rounded or estimated".
The vendor's own legal text says what the number cannot: "Codequiry does not determine whether plagiarism, academic misconduct or unauthorised use has occurred", and results "should be independently reviewed and interpreted by qualified personnel". A rate the vendor calls "low, not zero" lands on a real student every time it happens, which is why the same page asks for a human to read the flagged lines before anyone acts.
What independent studies found
No study record in the index has tested Codequiry, and none of the nine records in the studies index scored source code at all; every indexed sample is prose. Orenstrakh 2023, a computing-education paper that tested eight public detectors on student programming work, is on the candidate list and will be read in the primary source before any figure from it reaches this card.
The 1.3% figure above is therefore the only false positive number on the card, and it is the vendor's. The false positive ranking lists vendor-stated rates in a separate, unranked table for that reason.
Where Codequiry appears in the index
Rankings
- AIDI-R-07AI image, voice, video and code detectors ranked by independent studies
- AIDI-R-11AI detectors ranked by named integration surfaces
- AIDI-R-15AI detectors ranked by the LMS platforms the vendor names
- AIDI-R-17AI detectors ranked by the generator models the vendor names
- AIDI-R-19AI code detectors ranked by the programming languages the vendor names
- AIDI-R-20AI detectors ranked by launch year, with every change of hands and withdrawal the cards record
No ranking lists Codequiry. The card's false positive figure is vendor-stated, which the ranking rules leave unranked; the study count is zero; the free plan excludes AI detection; and the language field counts natural languages, which the vendor does not state (detectors.json, read 9 November 2026).
Codequiry appears on AI code detectors in the index, the audience view it was added for, as the one card built for code alone, and on AI detectors by content type, where code is a column. Every detector card on the register shows the same fields side by side, so a reader can set Codequiry's per-check metering beside the per-word credits of the text detectors.
Questions about Codequiry AI code detector
How much does Codequiry cost?
Grow is shown at C$272 a month, which the page's currency script derives from a US$200 base, with add-ons of C$406.64, C$950.64 and C$40.80 a month, all read 9 November 2026 on codequiry.com/pricing. The free plan covers peer checks only.
Which AI coding tools does Codequiry detect?
ChatGPT, GitHub Copilot, Claude, Grok, Google Gemini, Cursor AI, Amazon Q Developer and Meta Llama, with stated detection rates from 80% to 91%, plus a general model, per the detection page read 9 November 2026.
Does Codequiry keep the code students submit?
Yes, on every plan without the Rapid add-on: the privacy policy of October 22, 2026 says submitted code joins an anonymised cross-account corpus indefinitely and cannot be deleted on request. Rapid customers can delete submissions, read 9 November 2026.
Is there an independent test of Codequiry?
No record in this index as of 9 November 2026. The vendor's 1.3% false flag claim carries no sample size, and no indexed study has scored source code with any detector.
Sources and revisions
Sources
- codequiry.com/detect-ai-written-codeRead 9 November 2026
- codequiry.com/pricingRead 9 November 2026
- codequiry.com/aboutRead 9 November 2026
- codequiry.com/privacy-policyRead 9 November 2026
- codequiry.com/usage/apiRead 9 November 2026
