Copyleaks AI detector
Vendor Copyleaks Official site copyleaks.com/ai-content-detector How this card is sourced
Updated · AI Detector Index editors
Copyleaks AI Detector scans text, code, images and video and states support for 30+ languages, 11 of them named, according to the product page read on 8 November 2026 (copyleaks.com). Copyleaks is one of two detectors in the index to classify every document correctly in Walters 2023. How this card is sourced explains the rules; the full detector index holds the other twenty-one cards.
What Copyleaks detects
| Field | Value | Source |
|---|---|---|
| Content types detected | Text, code, image, video | copyleaks.com/ai-content-detector |
| Generators covered | GPT-5, GPT-4, ChatGPT, Claude, Gemini, DeepSeek, Llama (Meta), Bloom, Rytr, Jasper, Image: OpenAI (DALL-E and GPT), Google (Imagen and Gemini), Midjourney, Stability AI (Stable Diffusion), Black Forest Labs (Flux) | copyleaks.com/ai-content-detector |
| Languages | English, French, German, Italian, Portuguese, Spanish, Japanese, Chinese (Simplified), Chinese (Traditional), Russian, Dutch (30 listed) | copyleaks.com/ai-content-detector |
| Input limits | AI Detector Web-Based Platform: minimum 255 characters, maximum 2,000 pages (no character max). Browser extension: minimum 350 characters, maximum 25,000 characters. Free scan up to 25,000 characters. 1 credit covers up to 250 words or 1 image. OCR files: gif, png, bmp, jpg, jpeg. Image files: JPG, JPEG, PNG, WEBP, BMP, HEIC, HEIF. | copyleaks.com/ai-content-detector |
| Free tier | Free for up to 25,000 characters per scan; every new user receives a limited number of free credits (count not published); a free account unlocks more monthly credits | copyleaks.com/ai-content-detector |
| API | Yes, docs.copyleaks.com/resources/sdks/python | docs.copyleaks.com/resources/sdks/python |
| Integrations | Canvas, D2L, Moodle, Blackboard, Schoology, Edsby, Sakai, Google Classroom, Browser extension (Chrome, Edge, Firefox), Google Docs Add-on, WordPress add-on, MCP access | copyleaks.com/ai-content-detector |
| Data retention | Privacy Policy (Last revised August 24, 2026): Copyleaks may use data for purposes such as fraud detection, billing and improving Copyleaks' services including by training Copyleaks' models; direct customers may opt out of model training by contacting customer support; a Hide Sensitive Data feature is available for text-based content. API page FAQ: Content submitted to the Copyleaks AI Detector API is not stored, shared, or used to train public AI models. Pricing page: Shared Data Hub, a free, secure library of millions of user-submitted documents. | copyleaks.com/privacy-policy |
| Ownership | Copyleaks (together with its affiliated companies); no parent company stated | copyleaks.com/privacy-policy |
| Independent studies | 9 records | Studies index |
Copyleaks covers four modalities: text, code, image and video. For text the product page names GPT-5, GPT-4, ChatGPT, Claude, Gemini, DeepSeek, Llama (Meta), Bloom, Rytr and Jasper; for images, OpenAI, Google, Midjourney, Stability AI and Black Forest Labs (Flux) (copyleaks.com/ai-content-detector, read 8 November 2026).
The 11 named languages are English, French, German, Italian, Portuguese, Spanish, Japanese, Chinese (Simplified), Chinese (Traditional), Russian and Dutch, inside a stated total of 30+ (copyleaks.com/ai-content-detector, read 8 November 2026). The index records the count as 30.
On humanized text, the methodology page describes an Extra Sensitive setting: "Our most sensitive model was designed to flag AI text that was put through a 'humanizer' or text spinner" (methodology page, read 8 November 2026). The vendor also says Grammarly's AI-based rewriting features may trigger detection while basic grammar and spell-check functions typically do not (copyleaks.com/ai-content-detector, read 8 November 2026).
Copyleaks owns the product. Integrations cover Canvas, D2L, Moodle, Blackboard, Schoology, Edsby, Sakai and Google Classroom, a browser extension, Google Docs and WordPress add-ons, MCP access and an API with a Python SDK (copyleaks.com/ai-content-detector, read 8 November 2026).
Limits and plans
| Plan | Price | Limits | Checked on |
|---|---|---|---|
| Personal Plan (monthly) | $16.99/month | 100 unified credits: scan up to 25,000 words OR 100 images; full access to AI detection and plagiarism detection for text and AI detection for images. | |
| Personal Plan (annual) | $13.99/month when billed annually | Page also shows a 1,200 unified credits (300,000 words OR 1,200 images) tier, presumably the annual Personal allowance; label not readable. | |
| Pro Plan (monthly) | $99.99/month | Includes 25 user seats; 1,000 unified credits: scan up to 250,000 words OR 1,000 images. | |
| Pro Plan (annual) | $74.99/month when billed annually | Page also shows a 12,000 unified credits (3,000,000 words OR 12,000 images) variant. | |
| Education and Enterprise | Custom. |
The web platform takes a minimum of 255 characters and up to 2,000 pages with no character maximum; the browser extension takes 350 to 25,000 characters. One credit covers up to 250 words or 1 image (copyleaks.com/ai-content-detector, read 8 November 2026). The methodology page gives 350 characters as the minimum; this card uses the FAQ figures, which separate the two surfaces.
The free check covers up to 25,000 characters per scan; new users receive a limited number of free credits, and a free account unlocks more monthly credits (copyleaks.com/ai-content-detector, read 8 November 2026).
Prices below come from the Copyleaks pricing page, read 8 November 2026; the page shows a dollar symbol with no currency named.
On retention, the privacy policy revised August 24, 2026 says Copyleaks may use data for improving its services "including by training Copyleaks' models", with an opt-out for direct customers (privacy policy, read 8 November 2026). The API page says content submitted to the API "is not stored, shared, or used to train public AI models" (API page, read 8 November 2026).
What the vendor claims
Methodology page (test date September 10, 2026; publish date September 28, 2026; model V11): QA human-only total accuracy 0.9970 (923 of 326,120 human texts flagged as AI). Product page: accuracy stated as over 99%, with an industry-low .03% false positive rate; Copyleaks leads the field in trusted detection.
Copyleaks states accuracy of "over 99%" and "an industry-low .03% false positive rate", adding that "Copyleaks leads the field in trusted detection" (copyleaks.com/ai-content-detector, read 8 November 2026). The methodology page, test date September 10, 2026, model V11, reports a human-only total accuracy of 0.9970, with 923 of 326,120 human texts flagged as AI (copyleaks.com/ai-detector/testing-methodology, read 8 November 2026).
On false positives, Copyleaks says in the methodology sensitivity table: Balanced (default) false positives 0.30% and false negatives 0.64%; Extra Safe 0.16% and 0.88%; Extra Sensitive 0.66% and 0.49% (copyleaks.com/ai-detector/testing-methodology, read 8 November 2026). The product page gives .03%; the vendor's methodology page gives 0.30%; this card uses the methodology figure.
Copyleaks also says: "Our models need a certain volume of text to determine AI presence accurately" (copyleaks.com/ai-content-detector, read 8 November 2026). A Copyleaks flag on a short passage carries less weight than the same flag on a full document, by the vendor's own account.
What independent studies found
- AIDI-S-0003Walters 20232023
Copyleaks: 100% of 126 documents correct, 0 incorrect, 0 uncertain (0 false positives in 42 human papers; 100% on both GPT-3.5 and GPT-4 papers); one of two detectors (with Turnitin) to classify every document. Tested June 25 to July 12, 2023.
- AIDI-S-0004Elkhatat 20232023
Copyleaks: all 15 GPT-3.5 paragraphs flagged; GPT-4 sensitivity 93% (14 of 15, the highest of the five tools); 1 false positive in 5 human controls (specificity 80%).
- AIDI-S-0005Perkins 20242024
Copyleaks: highest baseline mean accuracy of the seven tools, 64.8% (Table 7); 73.9% on non-manipulated AI text falling to 58.7% after adversarial techniques (drop 15.2%, Table 8); 5 false accusations in 10 human samples (FAS 50%, the highest); 41 of 104 AI samples undetected (UDC 39%, the lowest).
- AIDI-S-0009Chaka 20232023
Chaka 2023: Copyleaks classified 15 of 21 English chatbot responses as AI-generated and 6 as human-written, the most correct of the five tools (rank 1 of 5); on ChatGPT responses machine-translated into German, French and Spanish, 3 of 6, 5 of 6 and 6 of 6 were classified as AI-generated. Free version, tested 30 March to 2 April 2023.
- AIDI-S-0012Orenstrakh 20232023
Ranked first of eight at 99.12% on 114 human and 95.00% on 20 ChatGPT English coursework submissions, with 1 false positive in 114; accuracy on 10 QuillBot-paraphrased ChatGPT texts fell from 100% to 50%, and all 10 Spanish human papers were flagged as AI (Tables 6, 8, 9 and 12).
- AIDI-S-0015Van Vlasselaer 20262026
Scored 75% of 40 fully AI-generated papers under 20% and 25% between 20% and 40%, reached 30.0% accuracy on 40 hybrid and 22.5% on 40 humanised papers, and classified all 40 human papers correctly (Tables 4 and 5).
- AIDI-S-0018Pegoraro 20232023
True positive rate 22.9% on ChatGPT responses and true negative rate 92.1% on human responses in a 20-method sweep on a ChatGPT and human Q&A benchmark, March 2023 (Table I).
- AIDI-S-0023Kar 20242024
Read one 500-word ChatGPT 3.5 article as 100% AI and the same text after a QuillBot paraphrase as 0 AI; 100 after Grammarly and ChatGPT paraphrases (Table 1, single text).
- AIDI-S-0024Alshammari and Rao 20252025
Average AI score 98.8% on 49 DeepSeek-V3 answers and average human score 100% on 49 human answers; 99.7% on DeepThink text, falling to 71% after a humanized paraphrase (Tables III and V, March 2025).
Walters 2023 found Copyleaks 100% correct on 126 documents between June 25 and July 12, 2023, with 0 false positives in 42 human papers; Copyleaks and Turnitin were the only two detectors to classify every document (De Gruyter PDF, read 8 November 2026).
Elkhatat 2023 reports that Copyleaks flagged all 15 GPT-3.5 paragraphs and 14 of 15 GPT-4 paragraphs (93%, the highest of five tools), with 1 false positive in 5 human controls (specificity 80%) (Springer PDF, read 8 November 2026).
Perkins 2024 gives Copyleaks the highest baseline mean accuracy of seven tools, 64.8%, falling from 73.9% on non-manipulated AI text to 58.7% after adversarial techniques, with 5 false accusations in 10 human samples, the highest, and 41 of 104 AI samples undetected, the lowest (arXiv 2403.19148, read 8 November 2026).
Chaka 2023 ran the free version of Copyleaks on 21 English chatbot responses between 30 March and 2 April 2023 and reports 15 classified as AI-generated and 6 as human-written, the most correct of the five tools (rank 1 of 5); on ChatGPT responses machine-translated into German, French and Spanish, 3 of 6, 5 of 6 and 6 of 6 were classified as AI-generated (DOI 10.37074/jalt.2023.6.2.12, read 9 November 2026).
Two records were added on 9 November 2026. Orenstrakh 2023 ranked Copyleaks first of eight at 99.12% on 114 human and 95.00% on 20 ChatGPT English coursework submissions, with 1 false positive in 114, a fall from 100% to 50% on 10 QuillBot-paraphrased texts, and all 10 Spanish human papers flagged as AI (arXiv 2307.07411, read 9 November 2026). Van Vlasselaer 2026 reports 75% of 40 fully AI-generated papers scored under 20% and 25% between 20% and 40%, 30.0% accuracy on 40 hybrid and 22.5% on 40 humanised papers, and all 40 human papers classified correctly (DOI 10.1007/s40979-026-00226-w, read 9 November 2026).
Pegoraro 2023, added 9 November 2026 as the seventh record, ran 20 detection methods on a ChatGPT and human question-and-answer benchmark in March 2023 and records Copyleaks at a 22.9% true positive rate on ChatGPT responses and a 92.1% true negative rate on human responses (arXiv 2304.01487, Table I, read 9 November 2026); no ranking field takes a value from the printed rates.
Two records joined the list on 9 November 2026 and took the count to 9: Kar 2024 read one 500-word ChatGPT 3.5 article at 100% AI and the same text at 0 AI after a QuillBot paraphrase, and Alshammari and Rao 2025 recorded a mean AI score of 98.8% on 49 DeepSeek-V3 answers and 71% after a humanized paraphrase; both print scores rather than rates, so no rank field moved.
Where Copyleaks appears in the index
Rankings
- AIDI-R-01AI detectors ranked by independent testing Two records added in the fifth wave on 9 November 2026, Pegoraro 2023 and Fleckenstein 2024, changed nine counts and none: Pegoraro 2023 tested ZeroGPT, GPTZero, Copyleaks, Originality.ai, Writer, Content at Scale (BrandWell), DetectGPT, the GPT-2 Output Detector and GLTR, lifting GPTZero to 12, the GPT-2 Output Detector to 10, Copyleaks to 7 and Originality.ai and ZeroGPT to 6; Fleckenstein 2024 tested 289 teachers and no software, so it counts for no card ([Fleckenstein 2024 record](/studies/fleckenstein-2024-teachers/)). Illuminarty, one of four cards added in the same wave, joined with its one image record from Ha and Passananti 2024, so 18 of 72 cards now qualify.
- AIDI-R-02AI detector false positive rates, ranked
- AIDI-R-04AI detectors ranked by languages supported
- AIDI-R-07AI image, voice, video and code detectors ranked by independent studies
- AIDI-R-08AI detectors ranked by published robustness to paraphrased and edited text
- AIDI-R-09AI detector false negative rates, ranked
- AIDI-R-10AI detector accuracy claims vs published results, ranked by the gap
- AIDI-R-11AI detectors ranked by named integration surfaces
- AIDI-R-12AI detectors ranked by the upload formats the vendor states
- AIDI-R-13AI detectors ranked by minimum text length per check
- AIDI-R-14AI detector APIs ranked by published price per 1,000 words
- 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-18AI detectors ranked by study records on non-English and non-native text
Fits these use cases, questions and comparisons
- AIDI-U-01AI detectors for business and managers
- AIDI-U-02AI detectors for school districts
- AIDI-U-04AI detectors for job applications and recruiters
- AIDI-U-05AI code detectors in the index
- AIDI-U-06AI detectors for admissions offices
- AIDI-U-08AI detectors for researchers checking their own manuscripts
- AIDI-U-09AI detectors for law firms and courts: an index fit check
- AIDI-U-10AI detectors for grant and proposal reviewers
- AIDI-U-11AI text detectors for newsrooms and fact desks
- AIDI-U-12AI detectors for localisation and translation teams
- AIDI-Q-01Is there a 100% accurate AI detector? What published tests show
- AIDI-Q-15Which AI detectors support Spanish? The cards that name it and the one study that tested it
- AIDI-Q-40Which AI detectors are free for nonprofits or education? The 12 cards that sell an institutional programme, the one university platform that charges nothing, and the two nonprofit tools that closed
- AIDI-C-13Pangram vs Copyleaks AI detector: two text detectors with one shared study, on the same 14 fields
Copyleaks qualifies for three of the five launch rankings and three of the five added on 9 November 2026: independently tested on 6 studies, false positive rate on the Walters 2023 figure of 0 in 42, and languages on a count of 30 (detectors.json, read 8 November 2026). The free allowance is per scan, not per month, so the card is absent from the free tiers ranking.
Since 9 November 2026 Copyleaks also ranks first in false negative rate on 0 of 84 AI papers missed in Walters 2023, fifth in robustness to paraphrase on 58.7% after the Perkins 2024 edits, and third in the claim vs evidence gap on "over 99%" claimed against 100% of 126 in Walters 2023; the card is listed in studies by content type with no record on its code, image or video detection.
Copyleaks fits business through the API and the 25-seat Pro plan, school districts through eight LMS integrations, agencies through the API and credit pools, recruiters through the free 25,000-character check, and code through the code modality (copyleaks.com/ai-content-detector, read 8 November 2026). The other code-capable text detector is Originality.ai; the closest LMS peer is GPTZero. The home page carries the index of AI detectors.
The card also sits in the two rankings added on 9 November 2026: minimum input at 255 characters and API pricing at $0.68 per 1,000 words from the Personal plan's credits.
Since 9 November 2026 Copyleaks also appears on five question pages added that day, among them the GPT-4 question (five records with GPT-4 family text), the false positive studies question (6 of its 9 records print a human-text figure) and the Canvas without Turnitin question, and ranks second on LMS coverage with six named platforms, sixth on generators covered with 15 named generators, and third on non-English evidence with three records.
Pangram vs Copyleaks, added 10 November 2026, reads this card's nine records beside Pangram's four, with Van Vlasselaer 2026 as the one record that ran both and the two vendor false positive figures, .03% and 0.30%, printed as the pages state them.
Questions about Copyleaks AI detector
What is the Copyleaks false positive rate?
0.30% at the default Balanced setting, per the vendor's methodology page dated September 10, 2026, read 8 November 2026. The product page says .03%. Walters 2023 recorded 0 false positives in 42 papers; Perkins 2024 recorded 5 in 10.
How much does Copyleaks cost?
$16.99/month for the Personal Plan billed monthly, or $13.99/month billed annually, for 100 unified credits, per the pricing page read 8 November 2026. The Pro Plan is $99.99/month monthly or $74.99/month annually.
Does Copyleaks detect AI-written code?
Yes. Code is one of four modalities on the product page read 8 November 2026, alongside text, image and video. The three studies on this card on 8 November 2026, and all six after the records added on 9 November 2026, cover text only.
Sources and revisions
Sources
- copyleaks.com/ai-content-detectorRead 8 November 2026
- copyleaks.com/pricingRead 8 November 2026
- copyleaks.com/ai-detector/testing-methodo...Read 8 November 2026
- copyleaks.com/api/ai-detectorRead 8 November 2026
- copyleaks.com/privacy-policyRead 8 November 2026
- docs.copyleaks.com/resources/sdks/pythonRead 8 November 2026
- degruyterbrill.com/document/doi/10.1515/o...Read 8 November 2026
- link.springer.com/content/pdf/10.1007/s40...Read 8 November 2026
- arxiv.org/pdf/2403.19148Read 8 November 2026
- doi.org/10.37074/jalt.2023.6.2.12Read 9 November 2026
Revision log
- Card created
- Chaka 2023 record added, 4 study records
- Orenstrakh 2023 and van vlasselaer 2026 records added, 6 study records
- Pegoraro 2023 record added to the study list in the fifth wave2026-11-09: kar 2024 and alshammari and rao 2025 records added in the sixth wave, 9 study records2026-11-09: seventh-wave question pages and three rankings linked2026-11-10: pangram vs copyleaks comparison linked (tenth wave)
). Illuminarty, one of four cards added in the same wave, joined with its one image record from Ha and Passananti 2024, so 18 of 72 cards now qualify.: register plate](/img/independently-tested-480.webp)
