Isgen AI is a multi-lingual AI detector that incorporated in February 2024[1]. If boasts detection services in over 80 languages while providing humanizing, plagiarism checking, and other standard AI writing services.
Pros
- Good plagiarism checker
- Can detects AI in multiple languages
Cons
- 1/7 against LLMs
- 2/8 against humanizers
Features
- AI detector: Identifies non-human writing in multiple languages.
- Plagiarism checker: Spots writing copied from another source.
- AI humanizer: Humanizes AI-written content.
- Bulk scan: Scans large quantities of documents for AI.
- AI image detector: Scans images to spot AI.
- Grammar checker: Uses AI to check spelling and grammar.
- API
Pricing Plans
| Trial | Academic | Writing | Ultimate |
| $0 | $14/month or $96/year | $25/month or $180/year | $40/month or $288/year |
| Limited AI detection, plagiarism checker, grammar checker | Includes AI detection, plagiarism checker, batch uploads | Includes grammar checker, humanizer, batch uploads, priority support | Includes all tools from Academic and Writing plans |
| Up to 6k characters per detection, up to 3k characters humanizing | 350k detection words/month | 50k humanized words/month | Unlimited words/month |
Test #1: Isgen AI vs. LLMs
For the first test, we asked the most widely-used AI LLMs to create a short article on the war of 1812. Each LLMs text was then run through Isgen to test it’s detection capabilities. The percentages below represent the likelihood of AI, not the amount of AI.
Prompt Used
Produce an article about the war of 1812 that reads as human as possible. It should be approximately 500 words long and be able to pass any AI detector. Please remove all instances of em dashes in the content.
Expand
Collapse
| LLM | Test Result | Proof |
| ChatGPT – 5.4 Thinking Model | 30% AI
❌ |
 |
| Claude – Sonnet 4.6 Extended |
0% AI ❌ |
 |
| Google Gemini – Thinking | 44% AI
🟡 |
 |
| Deepseek – Deepthink | 0% AI
❌ |
 |
| Perplexity – Free Model |
84% AI ✅ |
 |
| Grok – Auto |
26% AI ❌ |
 |
| Meta | 0% AI
❌ |
 |
Test #2: Isgen AI vs. Humanizers
We took the one successfully-detected AI article above (from Perplexity) and ran it through eight different online humanizers. The humanized result from each was in turn run through Isgen to again see if AI could be detected.
Expand
| AI Humanizer | Test Result | Proof |
| Humanize.ai (humanized text) |
32% AI
❌ |
 |
| StealthGPT (humanized text) |
0% AI
❌ |
 |
| WriteHuman.ai (humanized text) |
0% AI
❌ |
 |
| Undetectable.ai (humanized text) |
2% AI
❌ |
 |
| StealthWriter.ai (humanized text) |
0% AI
❌ |
 |
| WalterWrites.ai (humanized text) |
71% AI
✅ |
 |
| Grammarly (humanized text) |
100% AI
✅ |
 |
| Quillbot (humanized text) |
0% AI
❌ |
 |
Test #3: False-Positives + Plagiarism
We also checked Isgen to see if it would detect AI writing where there isn’t any and whether it could recognize plagiarized writing.
Expand
| Article URL | Test Result | AI Detection | Plagiarism |
| Harry Potter and the Goblet of Fire (excerpt) | 0% AI – 100% Plagiarized
✅ |
 |
 |
| President Obama Farewell Address (speech) | 0% AI – 100% Plagiarized
✅ |
 |
 |
| Wikipedia (“Netherlands”) | 0% AI – 100% Plagiarized
✅ |
 |
 |
| Newsweek (article) | 100% AI – 100% Plagiarized
❌ |
 |
 |
| BBC (article) | 0% AI – 100% Plagiarized
✅ |
 |
 |
| New York Post (article) | 32% AI – 100% Plagiarized
✅ |
 |
 |
| CNN (article) | 0% AI – 97% Plagiarized
✅ |
 |
 |
| NPR (article) | 0% AI – 100% Plagiarized
✅ |
 |
 |
| Yale School of Public Health (article) | 77% AI, 97% Plagiarized
❌ |
 |
 |
| Oxford Academic (article) | 19% AI – 100% Plagiarized
✅ |
 |
 |
Is Isgen.AI Accurate?
Isgen has an unclaimed profile on Trustpilot with only eight reviews, so it’s hardly conclusive. That said, 6 out of 8 reviewers gave it 1 star.
The results from our tests have us drawing the same conclusion as those reviewers on Trustpilot; Isgen.AI is not accurate.
Frequently Asked Questions
Do unused words carry over?
No, each month users words are refilled but unused words do not carry over.[2]
Are sentence-by-sentence breakdowns provided?
Yes, using Isgen’s detailed analysis, users can see the “impact” of each sentence on a scale of 1-5 (see image).
Is document scanning included?
Yes, documents can be uploaded in PDF, doc, and docx formats.[3]
Is image scanning provided?
Yes, users can upload images to check if they were generated by AI.[4]
Are reports shareable?
No, reports are not shareable directly from Isgen.ai, but a downloadable PDF report is provided with each scan.
Does Isgen.AI provide certifications?
No, while a generated report provides users with a shareable probability result, there’s no language that certifies the scan.
Are AI detections used for internal training?
There is no language anywhere on the Isgen.AI website that states definitively whether or not AI detections are used for internal training.
Does Isgen.AI store user data?
On their homepage,[5] Isgen states that they do not store user data and that input history can be deleted. However, in their privacy, they state clearly that they collect and use personal data to “provide and improve the Service.”[6]