StealthGPT – An In-Depth Review

StealthGPT, launched in November 2022, has become one of the internet’s most popular text “humanizers.” A recent study by the StealthGPT team has stated that the service is the top undetectable AI tool on the market. We were curious to see if StealthGPT could beat our AI detector tool of choice, 

After some testing, we have determined that StealthGPT is not perfect, but can get the job done in certain situations. We would not recommend using the service unless you are looking for an advanced supplementary tool. Although the output is well-written, the AI detector can still flag some sort of AI activity more than 60% of the time.

A Quick Overview

StealthGPT claims to be able to take your AI-generated content and turn it into something that will fool AI detectors. The service will cost you anywhere from $14.99-$29.99 per month, with the option of adding the “Stealth Samurai” for $4.99 more.

As far as the base plans go, the price difference is directly related to the number of words you can run per instance (we paid for the essential plan and received a cap of 750 words), and how many words you can run per month. See below for a quick breakdown of each plan:

  • Essential ($14.99/month) - 100,000 words/month, process up to 750 words
  • Pro ($19.99/month) - 500,000 words/month, process up to 1,500 words
  • Exclusive ($29.99/month) - 1,000,000 words/month, process up to 2,000 words

The price tiers seem fair and are explained well on the site. What we take issue with is the Stealth Samurai Engine. This $4.99/month add-on seems to just make the service better, something that should be included with the normal offering. The Stealth Samurai Engine upgrade gives users increased coherency of engine output and claims to beat even more AI detectors (like Winston AI and CopyLeaks).

The UI is clean and intuitive, making the process easy for anyone looking to transform their AI content. The app even gives the user four different choices on how they would like their writing to feel, with a standard, casual, professional, and academic setting. 


Once your input is processed, the app spits out a new and improved undetectable version. I will say that compared to most text “humanizers,” the StealthGPT app does a great job of keeping the article intact. Many text “humanizers” resort to adding poor grammar, misplaced punctuation, and awkward phrasing; something that hinders the user even more when trying to create quality content. 

The output looks something like the image below; all you have to do is copy the text and paste it into a document. 

Stealth Output

One complaint that is worth noting is in the processing word limit itself. We ran a couple of articles over the word limit just to see what would happen. The app told us that we were over the word limit, but still docked us in our monthly word count. This is not a problem if you are carefully processing under the word limit every time, but if you have a lot of articles to run you may experience the same issue.

StealthGPT vs Originality

At BrandBuilders, we have used as our go-to plagiarism and AI detection tool. Every article we publish is run through Originality first, and we have noticed great accuracy with the service. We wanted to test StealthGPT against Originality, to see if it can stack up against one of the best tools on the market.

We had ChatGPT generate 10 separate articles on varying topics, all ranging from 600-1,000 words. Then, we used those AI-generated articles and ran them through StealthGPT’s app, which produced 10 new articles poised to take on Originality. Our results (seen below) showed us that even though StealthGPT works occasionally, it would not be reliable enough for someone looking to consistently beat AI detectors.

StealthGPT vs Originality

As you can see, when the ChatGPT articles were run through Originality, they (expectedly) came back with a 100% AI Detection score. When we ran the StealthGPT doctored articles through, only 4 out of 10 came back with overwhelming Original scores. As an editor, anything scoring over 25% would spark suspicion. The distribution chart below shows Originality’s confidence that an article was written by AI.

StealthGPT Vs Originality AI Score analysis

Our Verdict

StealthGPT is very clear about its offering, and users should not expect it to work 100% of the time. But it is slightly concerning that the app struggled in our limited testing. 60% of our tested articles would raise some sort of suspicion from an editor, teacher, or anyone else who could be scanning your work.

The AI arms race is heating up, and the dynamic between text “humanizers” and AI detectors is an ever-evolving situation. As for now, it seems like Originality can still detect even the most sophisticated software. StealthGPT is one of the best software options we have tested, but its inclusion of the Stealth Samurai Engine and lack of consistency may be off-putting for some.