isgen.ai checks whether text was written by a person, an AI model or both, with sentence- and phrase-level scores in over 80 languages. The suite also covers plagiarism, image and video detection, and sells an AI humanizer.
Used this tool? Rate it
Used this tool? Rate it
isgen.ai (also branded Isgen) is a web-based AI content detector that estimates whether a text was written by a person, generated by a language model, or produced by a mix of both. It is positioned as an AI detector for text written by GPT-5, ChatGPT, Claude, Gemini, DeepSeek and other open-source models. Around that detector sits a wider suite: a plagiarism checker, grammar and citation tools, an AI editor, separate image and video detectors, and an AI humanizer.
Two facts shape most decisions about it. The headline accuracy figures are published by the vendor itself, and the same pages say the detector is not 100% accurate. And the suite sells a humanizer marketed for passing AI detectors next to a detector marketed to teachers, which matters for anyone using a score to judge someone else's work.
The service is run from the UK. The terms of service name ISGEN LTD, a company registered in England at 71-75 Shelton Street, London. The privacy policy instead names ISGEN LLC at the same London address, so the two legal documents do not use the same entity name. A third-party traffic estimate put isgen.ai at about 703.6K monthly visits for May–August 2026, as captured on September 28, 2026.
The AI content detector is the centre of isgen.ai, and most other tools either feed text into it or act on its results.
The human-leaning default and the sentence and phrase views work together. Instead of a single percentage, a reviewer can see which passages push the score up, while borderline passages are resolved towards human-written rather than AI.
The AI Humanizer rewrites text for a more natural tone and is sold by the same company as the detector. Two writing styles are offered: Formal for a professional tone and Adaptive for keeping the writer's own voice. It has a Basic mode for surface-level changes and an Advanced mode that the product page advertises as delivering "100% AI bypass". The same page states that use for academic misconduct is strictly prohibited and that isgen.ai is not responsible for misuse. The humanizer is described as supporting more than 60 languages, fewer than the 80-plus claimed for the detector.
isgen.ai publishes its own guidance for educators who find AI-flagged text in student work. The first step it describes is to open the Deep and Detailed Insights view and look at which parts of the text have the highest impact score, since those parts may be what is pushing the overall AI score up. The second is to let the student explain how the work was produced, for example through an in-person discussion, notes on how the ideas were developed, or the version history of the editor the student used. The same guidance points to the institution's own rules on permitted AI use as the boundary.
isgen.ai runs in a web browser; no mobile app or browser extension was found in the pages checked.
A free AI detector is available without payment. Its text box shows a 5,000-character counter. With yearly billing selected, the pricing page shows Academic at $8/month, Ultimate at $24/month and Writing at $15/month, each billed annually. A limited-time banner advertises the annual option as "Our Biggest Annual Offer: Get 5 Months Free", and monthly-billing prices were not displayed in the captured default view.
| Plan | Monthly word allowance | AI detector per submission | AI Humanizer | Bulk AI scan |
|---|---|---|---|---|
| Academic | 350,000 detection words | 60,000 characters | Not included | 100 |
| Ultimate | Unlimited | Unlimited | Included | 150 |
| Writing | 50,000 humanizer words | Not included | Included | Not included |
In the comparison table, Writing lists the AI detector as not included even though the same table shows a 6,000-character detector limit for that plan. Words do not roll over to the next month. Paid users can buy additional words from the account page. Upgrades are charged pro rata for the rest of the current billing cycle.
Subscriptions renew automatically until cancelled. Cancelling keeps premium access and remaining words until the end of the billing cycle, after which the account drops to the Freemium plan. The terms list Visa, Mastercard, American Express, Discover, PayPal, Link and Apple Pay as accepted payment methods.
The refund rules differ between two isgen.ai pages:
Copyleaks and GPTZero both appear among the related AI detectors listed on a third-party review-platform profile for isgen.ai, and each differs on a point that can decide the choice.
In short, isgen.ai's case rests on language breadth and sentence-level insight, while the two alternatives put more weight on school-system and developer integrations.
No independent study or media report covering isgen.ai was found in the searches made for this page, so the accuracy and false-positive figures rest on the vendor's statements alone. That matters most for AI detector false positives in education, where a score can affect a student's record.
The humanizer page is titled "AI Humanizer Free: Pass AI Detectors & Humanize AI Text", while the detector from the same company is promoted to educators and students. The writer use case above, checking that copy will not be flagged as AI-generated, sits between the two products.
The privacy policy is dated February 10, 2024, and in the pages checked it does not address the text or files submitted for detection, nor whether they are used to train models. The only training statement found is the one on the bulk scan page.
Yes, within limits. The text detector can be used free with a 5,000-character input box, and the image detector is advertised as free with unlimited scans. Paid plans start at $8/month billed annually and add higher word allowances and bulk scans.
The company reports 96.4% accuracy on its own benchmark and a false positive ratio of nearly 0%, but it also says no AI detector is 100% accurate. No independent study of isgen.ai was found, so these remain vendor claims.
The statements differ by page. The homepage FAQ says no data is stored, yet paid plans list 15 or 30 days of document history, and the bulk scan page says uploaded files are deleted after processing and are not used for training.
Possibly, but the rules conflict. The pricing FAQ ties eligibility to using less than 15% of the plan, while the refund policy requires a request within 12 hours of first subscribing and under 10,000 words used; yearly plans are refunded at 50%.
Yes. The text detector is described as covering over 80 languages and detecting the language automatically, while the humanizer covers more than 60 languages.
