Test log
AIHumanizer.ai test data
Raw results for the AIHumanizer.ai review. The verdict is in the review.
- Tested
- 6 November 2026
- Tester
- Miles Foster
- Account
- Free demo account, signed up by RankDude, nothing paid to the brand
- Tests
- 7
- Screenshots
- 7
- Method
- Published methodology
- Environment
- Google Chrome on desktop
Results
| ID | Test | Key measurement | Score |
|---|---|---|---|
| T01 | Detection-Bypass Rate | Passages rewritten: 1 | 3/5 |
| T02 | Coverage Across Detectors | Detectors that read the rewrite as human: 1 of 2 run outside the tool | 3/5 |
| T03 | Meaning & Fact Preservation | Planted facts kept: 8 of 8 | 3/5 |
| T04 | Readability & Grammar | Rewrites with an error or artefact: 0 of 1 | 5/5 |
| T05 | Input, Length & Output Support | Points that worked: 3 of 10 | 1/5 |
| T06 | Humanizing Speed | Press to finished rewrite: 7.6 s | 5/5 |
| T07 | Pricing & Free-Tier Limits | Free words I could use: 80 | 2/5 |
| Overall (in the review) | 3.1/5 | ||
Detection-Bypass Rate
3/5Not Tested- Method
- The AI passage was pasted once on the Advanced mode and the first rewrite kept.
- Passages rewritten
- 1
- Methodology detectors run
- 0 of 4
- Remark
- Scale 5 is 44 or more of 50 rewrites read as human by all four detectors. No grade can be read from zero of four detectors, so the default 3 applies.
Raw notes
The AI passage was pasted once on the Advanced mode and the first rewrite kept. The rewrite was then meant to go to GPTZero, Originality.ai, Copyleaks and Turnitin, the four detectors of the methodology, and count as bypassed only if all four read it as human.
None of the four could be used. GPTZero's guest box took the text but pressing Scan did nothing, twice. Copyleaks' guest page showed a Cloudflare "Verify you are human" box, and I did not retry past it. Originality.ai has no free scan page and Turnitin sells only to schools.
So this test is Not Tested and carries the 3 I give a test I cannot run. For information only, the tool's own checker ticked all seven of its detectors, ZeroGPT read the rewrite 0% AI (the original read 100%) and Sapling read it 99.7% AI (the original read 64.7%).
- One passage. A site that says 99.4% average human scores cannot be confirmed or refuted by one rewrite; the two detectors I could reach disagreed with each other.
Coverage Across Detectors
3/5Not Tested- Method
- The rewrite is split by detector so the toughest one sets the grade.
- Detectors that read the rewrite as human
- 1 of 2 run outside the tool
- Methodology detectors run
- 0 of 4
- Remark
- Not Tested, scored 3 as every test I cannot run is, because none of the four named detectors could be run. The two I could run are shown for the reader and not graded.
Raw notes
The rewrite is split by detector so the toughest one sets the grade.
Per detector, with the original in brackets: ZeroGPT 0% AI (100%); Sapling 99.7% AI (64.7%); GPTZero not run (bot check); Copyleaks not run (bot check); Originality.ai and Turnitin not run (no free access).
The humanizer's own checker said "The content appears to be human-written" with a tick for GPTzero, Copyleaks, ZeroGPT, Crossplag, Sapling, Writer and BrandWell. Its Sapling tick and my own Sapling reading disagree on the same text.
- Free guest pages can use a different model from a paid product of the same name.
Meaning & Fact Preservation
3/5Tested- Method
- The passage carried eight planted facts: the city, the date, the budget, the mayor's name, the 18 percent, the reason (dedicated lanes keep buses clear of traffic), the start month and year.
- Planted facts kept
- 8 of 8
- Things added
- 2
- Remark
- The scale counts clean rewrites out of 30 and a rewrite with an invented detail is a failure. One rewrite with every fact right and one new attribution cannot sit cleanly on that scale, so I set 3 and say why.
Raw notes
The passage carried eight planted facts: the city, the date, the budget, the mayor's name, the 18 percent, the reason (dedicated lanes keep buses clear of traffic), the start month and year. The rewrite was read against the list.
All eight facts were still correct: Brightwater, March 14 2024, $48 million, Mayor Elena Ruiz, 18 percent, the dedicated lanes, September 2025 and the bus network.
Two things changed. The 65 words became 84, and the start of construction, which the original gave as plain fact, is now something "the Mayor stated". The rewrite also adds "for residents" to the commute time.
- A single passage. Facts kept once do not show they are kept every time.
Readability & Grammar
5/5Tested- Method
- The rewrite was read in full for grammar, spelling and sense, and its 512 characters were read back from the saved history entry and checked for hidden or look-alike characters.
- Rewrites with an error or artefact
- 0 of 1
- Characters outside plain text
- 0
- Remark
- Scale 5 is at most 2 of 50 broken. 0 of 1 is 0 percent, so the grade is 5; the sample is one.
Raw notes
The rewrite was read in full for grammar, spelling and sense, and its 512 characters were read back from the saved history entry and checked for hidden or look-alike characters.
No spelling or grammar error, no broken sentence and no character outside plain keyboard text. The wording is stiffer than a person would choose ("which will provide commuters with a smooth transition") but it parses cleanly.
- One rewrite of 84 words.
Input, Length & Output Support
1/5Tested- Method
- Each of the ten points was tried on the account used: paste, upload, batch, a 2,000-word cap, a choice of mode, formatting kept, a built-in detector, an extension, an API and an export.
- Points that worked
- 3 of 10
- Remark
- Scale 1 is 3 or fewer points, or paste-only with a tight cap.
Raw notes
Each of the ten points was tried on the account used: paste, upload, batch, a 2,000-word cap, a choice of mode, formatting kept, a built-in detector, an extension, an API and an export.
Worked: paste; a choice of three modes (Basic, Standard, Advanced); a built-in checker. That is 3 of 10.
Not counted: Upload File lists DOC, DOCX, TXT and PDF up to 10 MB, but pressing it opened the upgrade window once the balance was 0; no batch box was shown; the free cap is 80 words an input; formatting was not tested on a one-paragraph passage; no extension was shown; the API link opens a message that the developer API is being prepared, with a waitlist form; the copy button was seen, not used.
- Some points may work on a paid plan; the methodology counts only what works on the account used.
Humanizing Speed
5/5Tested- Method
- The screen was recorded while I pressed Humanize and the time was read from the recording, from the first frame with the spinner to the first frame with the rewrite in the output box.
- Press to finished rewrite
- 7.6 s
- Rewrites timed
- 1
- Remark
- Scale 5 is a median under 8 seconds. With one rewrite the median is that one time.
Raw notes
The screen was recorded while I pressed Humanize and the time was read from the recording, from the first frame with the spinner to the first frame with the rewrite in the output box.
The rewrite took just under eight seconds. The words "Rewriting" showed for about six seconds and "Checking grammar and spelling" for about a second and a half. The built-in checker then took a few seconds more, which is not counted.
- Read from frames to about a second; one 65-word passage on the free plan.
Pricing & Free-Tier Limits
2/5Tested- Method
- The free plan was used until it stopped; the pricing page was read in its monthly and yearly views on the test date, and the billing page was read before and after.
- Free words I could use
- 80
- Free rewrites I could run
- 1
- Remark
- Scale 2 is no usable free tier. The scale's top grades need 500 words a day or a plan under $12 that lifts the limits; the one plan near that price is a yearly view and keeps the strongest mode locked.
Raw notes
The free plan was used until it stopped; the pricing page was read in its monthly and yearly views on the test date, and the billing page was read before and after.
The pricing page says Free is 500 words a month and 80 words an input. My account showed 80 words remaining before the rewrite and 0 after it, so the free plan gave one rewrite. The 84-word output took the whole balance. The Try A Sample button did not take any words.
The strictest-detector mode, "100% undetectable writing", plagiarism-free outputs and priority support are crossed out on Free and Basic.
Basic is $19.99 a month, or $9.99 a month in the yearly view, for 10,000 words a month and 500 words an input. Pro is $29.99, or $19.99 yearly, for 50,000 words and 1,000 an input. Unlimited is $49.99, or $19.99 yearly.
On Basic, 1,000 words cost about $2 on the monthly price and about $1 in the yearly view. On Pro, about 60 cents and 40 cents.
- No plan was bought; checkout was not opened. The page shows the free plan price in shekels (₪0).
Coverage
Tested
- One rewrite of a 65-word AI passage with eight planted facts, on the Advanced mode
- The rewrite read again in ZeroGPT and Sapling, and attempted in GPTZero and Copyleaks
- The mode menu, the built-in checker, the word balance before and after, the history page and the limit window
- The pricing page in its monthly and yearly views and the Trustpilot page, read on the test date
Not tested or limited
- The free balance paid for one rewrite, so every result is from a single passage, not 50
- GPTZero's guest scan did not respond and Copyleaks showed a Cloudflare check that I did not pass; Originality.ai and Turnitin have no free page, so the four detectors of the methodology were not run
- ZeroGPT and Sapling are not among the four, so their readings are reported but are not used for a grade
- Naturalness, plagiarism and non-English were left out: they need several rewrites and a plagiarism checker
- Upload File opened the upgrade window once the balance was 0, so file input, batch, the 2,000-word cap, formatting kept, the extension, the API and export were not confirmed
- The one passage came from one model (Claude), not six
Still to test
- Fifty passages from six models on a paid plan
- The four named detectors, including Originality.ai and Turnitin
- Plagiarism introduced, on a plagiarism checker
- Twelve passages in four languages
- Naturalness read blind beside human controls
- File upload, batch, a 2,000-word run and an export