RankDude

Test log

WriteHybrid test data

Raw results for the WriteHybrid 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
6
Environment
Google Chrome on desktop

Results

IDTestKey measurementScore
T01Detection-Bypass RatePassages rewritten: 43/5
T02Coverage Across DetectorsZeroGPT: rewrites read as human: 4 of 43/5
T03Meaning & Fact PreservationRewrites with every fact and meaning kept: 3 of 43/5
T04Readability & GrammarRewrites with an error or artefact: 1 of 43/5
T05Input, Length & Output SupportPoints that worked: 4 of 102/5
T06Humanizing SpeedMedian, press to finished rewrite: 13.7 s4/5
T07Pricing & Free-Tier LimitsFree words in the account: 5003/5
Overall (in the review)3/5
T01

Detection-Bypass Rate

3/5Not Tested
Method
Each AI passage was pasted once on the default settings and the first rewrite kept.
Passages rewritten
4
Methodology detectors run
0 of 4
Rewrites ZeroGPT read as human
4 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.

ZeroGPT reading the WriteHybrid rewrite of the Brightwater passage as human written with 0% AI GPT.

ZeroGPT reading the WriteHybrid rewrite of the Brightwater passage as human written with 0% AI GPT.
T01-1 ZeroGPT on the first rewrite: 0% AI.
Raw notes

Each AI passage was pasted once on the default settings and the first rewrite kept. The rewrites were 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 them as human.

None of the four could be used. GPTZero's guest box took the text but pressing Scan did nothing. Copyleaks' guest page left its Scan button disabled with the text box empty, and I did not push past it. Originality.ai has no free scan page and Turnitin sells only to schools.

So this test is Not Tested and counts as 3 by default. For information only, ZeroGPT read all four rewrites as human (the originals read 100%, 100%, 100% and 84.5% AI), and Sapling read two of the four as human.

  • Four passages. The site's own claim of bypassing GPTZero, Turnitin, Originality.ai and Copyleaks cannot be confirmed or refuted from four rewrites.
  • A cookie notice covered the Copyleaks page and I did not dismiss it, so its guest scan was never tried past that point.
T02

Coverage Across Detectors

3/5Not Tested
Method
The rewrites are split by detector so the toughest one sets the grade.
ZeroGPT: rewrites read as human
4 of 4
Sapling: rewrites read as human
2 of 4
Methodology detectors run
0 of 4
Remark
Not Tested, counted as 3 by default, because none of the four named detectors could be run. The two I could run are shown for the reader and not graded.

Sapling's AI detector scoring the WriteHybrid rewrite of the Brightwater passage at 86.7% AI with every sentence highlighted red.

Sapling's AI detector scoring the WriteHybrid rewrite of the Brightwater passage at 86.7% AI with every sentence highlighted red.
T02-1 Sapling on the first rewrite: 86.7% AI.
Raw notes

The rewrites are split by detector so the toughest one sets the grade.

Per detector, original then rewrite, for the four passages: ZeroGPT 100% then 0%, 100% then 32.7%, 100% then 0%, and 84.5% then 14.1%. Sapling 64.7% then 86.7%, 99.5% then 27.8%, 99.8% then 0.4%, and 84.0% then 63.4%.

With 50% as the line, ZeroGPT read 4 of 4 rewrites as human and Sapling read 2 of 4. On the first passage Sapling scored the rewrite higher than the original. GPTZero, Copyleaks, Originality.ai and Turnitin were not run.

The tool's own checker said "Your text is Human" with 99% and 6/6 detectors passed on every one of the four rewrites, including the two that Sapling read as AI.

  • Free guest pages can use a different model from a paid product of the same name.
T03

Meaning & Fact Preservation

3/5Tested
Method
Each passage carried planted facts: names, dates, figures and a claim with its reason.
Rewrites with every fact and meaning kept
3 of 4
Planted facts lost
0
Labels changed
1
Remark
The scale counts clean rewrites out of 30. Three of four is 75 percent, which sits in the 20 to 24 band, so the grade is 3.

WriteHybrid with the 65-word Brightwater original on the left and the rewrite on the right, changed words highlighted.

WriteHybrid with the 65-word Brightwater original on the left and the rewrite on the right, changed words highlighted.
T03-1 The original and the rewrite side by side.
Raw notes

Each passage carried planted facts: names, dates, figures and a claim with its reason. Each rewrite was read against its list.

Three of the four rewrites kept every fact and the argument. The Brightwater rewrite (65 words) kept the city, the 14 March 2024 date, the $48 million, Mayor Elena Ruiz, the 18 percent, the reason (designated lanes keep buses out of traffic) and the September 2025 start. The Valdoria rewrite kept the 42 kilometres, six districts, the 15 percent in two years and its reason, 8.5 million euros and the 9 percent rise. The Halvorsen rewrite kept June 2023, 6 percent, 18 schools, Dr. Priya Raman, 14 months and 3,200 students.

The email failed. Its "Subject:" line came back as "Name tag:", and "I hope this message finds you well" became "I hope you are dong well". The order, the 12 December ship date, the 240 chairs and the 1,150 dollars survived.

  • Four passages from one model. Facts kept four times do not show they are kept every time.
T04

Readability & Grammar

3/5Tested
Method
Every rewrite was read in full for grammar, spelling and sense, and each was checked for hidden or look-alike characters.
Rewrites with an error or artefact
1 of 4
Hidden or look-alike characters
0
Remark
Scale 3 is 7 to 12 broken out of 50 and scale 2 is 13 to 22. One of four is 25 percent, which is 12.5 out of 50, so I set 3 and say that it sits on the line.

WriteHybrid comparison window with the email original on the left and the humanized email on the right, with Name tag, dong and sincerly highlighted in green.

WriteHybrid comparison window with the email original on the left and the humanized email on the right, with Name tag, dong and sincerly highlighted in green.
T04-1 The email: original left, rewrite right.
Raw notes

Every rewrite was read in full for grammar, spelling and sense, and each was checked for hidden or look-alike characters.

One of the four was broken: the email carries "dong" for doing, "sincerly" for sincerely and "Name tag" for the subject line, and ends "Thanks for your patience and comprehension". The other three parse cleanly, though the first ends on "for convenient commuter connectivity" and the Halvorsen rewrite has double spaces after three sentences.

No rewrite holds a character outside plain keyboard text, apart from a curly apostrophe.

  • Four rewrites of 65 to 112 words.
T05

Input, Length & Output Support

2/5Tested
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
4 of 10
Remark
Scale 2 is 4 or 5 points.

WriteHybrid's settings panel with Purpose, Complexity Level and Number of Outputs blurred behind an Unlock Advanced Features button.

WriteHybrid's settings panel with Purpose, Complexity Level and Number of Outputs blurred behind an Unlock Advanced Features button.
T05-1 The settings panel on the free plan.
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 built-in detector panel under every rewrite; the copy button, which answered "Copied"; and an API, whose bearer key and docs sit on the free account (I did not call it, to keep the words). That is 4 of 10.

Not counted: no upload button on the humanizer; one passage at a time; the free cap is 500 words an input against 2,000 on Starter; the modes and complexity are locked behind "Unlock Advanced Features"; formatting was not tested because every passage was one paragraph; the Chrome extension is a zip file you load yourself, and I did not install it. The separate AI Detector page is locked to paid plans.

  • Some points may work on a paid plan; the methodology counts only what works on the account used.
  • Formatting through the rewrite and the Chrome extension were not tried; a retest with a multi-paragraph passage and the extension loaded could add up to two points.
T06

Humanizing Speed

4/5Tested
Method
The screen was recorded while I pressed Humanize and the time was read from press to the first moment the rewrite was on screen.
Median, press to finished rewrite
13.7 s
Fastest and slowest
12.4 s and 14.0 s
Rewrites timed
4
Remark
Scale 4 is a median from 8 to 14.9 seconds.
Raw notes

The screen was recorded while I pressed Humanize and the time was read from press to the first moment the rewrite was on screen.

The four rewrites took 13.6, 13.7, 14.0 and 12.4 seconds, a median of 13.7 seconds. A waiting window with changing messages and a button that offers faster processing covers the page until the text appears. The built-in checker then runs for a few more seconds, which is not counted.

  • A script read the page about four times a second from pressing Humanize to the first moment the rewrite was on screen, and the first run was checked against the screen recording; 65 to 112 words each, on the free plan, whose waiting window offers faster processing.
T07

Pricing & Free-Tier Limits

3/5Tested
Method
The free plan was used and the plans were read in the upgrade window in both billing views on the test date.
Free words in the account
500
Free words I used
350
Free words left
150
Remark
Scale 3 is a tight free tier, or a paid plan from $20 to $29.99. The free tier is tight: 500 words, one passage at a time, with the advanced settings and the detector page locked. The cheapest plan is under $20 and lifts the cap.

WriteHybrid's pricing section in the yearly view with the Starter, Pro and Agency plans and their word limits.

WriteHybrid's pricing section in the yearly view with the Starter, Pro and Agency plans and their word limits.
T07-1 The plans on 12 November 2026, yearly view.
Raw notes

The free plan was used and the plans were read in the upgrade window in both billing views on the test date.

The page title promises 500 free words monthly. My account held 500 words in total, and four passages of 350 words left 150. I used the allowance once, so I cannot say whether it renews.

Starter is $19 a month, or $16 a month billed yearly as $190, for 20,000 words a month and 2,000 words an input. Pro is $49, or $41 yearly, for 60,000 words and 5,000 an input. Agency is $99, or $83 yearly, for 150,000 words and 10,000 an input.

On Starter, 1,000 words cost about 95 cents on the monthly price and about 80 cents billed yearly.

  • No plan was bought; checkout was not opened.

Coverage

Tested

  • Four AI passages rewritten once each on the default settings, with their facts checked
  • Each rewrite read again in ZeroGPT and Sapling, and attempted in GPTZero and Copyleaks
  • The settings panel, the AI Detector page, the history list, the copy button, the account and API tabs and the extension window
  • The upgrade window in its monthly and yearly views and the Trustpilot page, read on the test date

Not tested or limited

  • The free balance of 500 words paid for four short passages, so every result comes from four rewrites, not 50
  • GPTZero's guest Scan did nothing and Copyleaks' Scan button stayed disabled; Originality.ai has no free page and Turnitin has no public one, 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 many rewrites and a plagiarism checker
  • The advanced settings are locked on the free plan, so modes, strength and several outputs were not tested; the Chrome extension was not installed and the API was not called, to keep the words
  • The passages 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
  • The advanced settings, a 2,000-word run and several outputs
What these results meanRead the full WriteHybrid review