Reduce AI Detection
Below 20% in our tests, with a 99% pass rate. Works with 8 detectors including CNKI and VIP, and rewrites from your detection report.
Supports .docx / .txt / .md, 80,000 characters per run. Credits are charged by actual usage. See credit rules
How to lower the AI rate of a paper
- 1
Submit the paper and the report
Paste the text or upload a file; if you have a CNKI, VIP or other detection report, upload it too so flagged passages are rewritten first.
- 2
Pick the output language
Chinese follows the Chinese de-templating skill; English follows the academic humanizer skill.
- 3
Rewrite paragraph by paragraph
Only filler, clichés and formulaic patterns change; terms, data, formulas and citations stay.
- 4
Re-test on the required platform
Use the detector your school requires, comparing before and after on the same platform and text range.
Lower the AI rate, keep the paper intact
Terms, data and meaning are kept, and edits focus on templated phrasing.
Lower AI rate, same rigor
Academic register, qualifiers and argument structure are kept; nothing is made colloquial. Terms, data, formulas, tables and references stay as they are.
Meaning and certainty stay the same
Correlation stays correlation, "may" stays "may", and no information is added or dropped; only filler with no meaning of its own is removed.
Targeted rewriting from your school's detection report
Works with 8 common detectors including CNKI and VIP. Upload your detection report and the flagged passages are rewritten; in our tests papers drop below 20% with a 99% pass rate. If the full-text AIGC share on CNKI or VIP is still above 25% after processing, contact support for a credit refund.
Only templated sentences change
The 31 checkpoints find empty, repetitive or formulaic sentences to rewrite, and fluent paragraphs stay as they are.
See an AI-trace reduction round, start to finish
From selecting the text to reverting just one paragraph, it happens step by step in a workspace. The sample was written to show the workflow.
With the rapid development of artificial intelligence technology, its application in education has become increasingly widespread, profoundly transforming traditional teaching models.Restored
Notably, this study found that students using intelligent tutoring systems improved their math scores by an average of 12.4 points (p < 0.05) [3].In this study, students using intelligent tutoring systems improved their math scores by an average of 12.4 points (p < 0.05) [3].
In conclusion, artificial intelligence can not only improve teaching efficiency but also promote personalized learning, demonstrating broad application prospects and significant research value.Artificial intelligence can help improve teaching efficiency and creates room for personalized learning.
Struck-through text is removed; shaded text is the new wording.
Submitting soon, keep the data and citations
Change notes
- FillerRemoved "Notably" and "In conclusion"
- PatternSplit the "not only... but also" construction
- HypeRemoved "broad prospects and significant value"
Kept as is：12.4 points, p < 0.05, citation [3]
Keep my original wording for paragraph 1
Paragraph 1 is restored; paragraphs 2 and 3 keep the rewrite.
Why it gets below 20% without distorting the paper
Chinese de-templating
Follows the 31 checkpoints of the open-source Humanizer-zh project to cut filler and break formulaic patterns, keeping facts, qualifiers and the author's voice.
Browse the skill marketAcademic humanizer for English
English papers and NSF or NIH proposals follow academic-humanizer: AI clichés and overclaims are removed, and every number and citation is kept.
Keep editing in the workspace
Every rewritten paragraph is visible in the editor. Revert one with a sentence, or keep adjusting only the paragraphs you choose.
Acknowledged in published SCI papers
Published papers state in their acknowledgments or AI-use statements that Acadwrite was used to polish the language.
- Theranostics2024Microfluidic chips in female reproduction: a systematic review of status, advances, and challenges
During the preparation of this work the authors used Acadwrite in order to improve readability, grammar and language style.
- Frontiers in Neuroscience2024Physical exercise for brain plasticity promotion an overview of the underlying oscillatory mechanism
We would like to thank Acadwrite (acadwrite.cn) for English language editing.
- Poultry Science2025Effects of induced molting on lipid accumulation in liver of aged laying hens
During the preparation of this work authors used (https://acadwrite.cn/) in order to improve language.
- Aging Medicine2025The Mechanism and Potential Therapeutic Strategies of Vascular Aging
Acadwrite AI (https://acadwrite.cn/apps/polish) was utilized for polishing the English language of the manuscript to enhance readability.
- Interactive Journal of Medical Research2025Evolution of Artificial Intelligence in Medical Education From 2000 to 2024: Bibliometric Analysis
During the preparation of this work, the authors used Acadwrite to improve readability, grammar, and language style.
AI-trace reduction FAQ
Still have questions? Rewriting is charged by actual usage; see the credit rules for details. See credit rules
What is the success rate? How low can it go?
In our tests, papers drop below a 20% AIGC rate with a 99% pass rate. We target CNKI and VIP; if the full-text AIGC share on either is still above 25% after processing, contact support for a credit refund.
Which detectors are supported? Can I upload a report?
It works with 8 common detectors including CNKI and VIP. Upload the detection report together with the paper and the flagged passages are rewritten first, while unflagged passages stay as they are.
How do I lower the AI rate of my paper?
Confirm which platform your school uses and get a baseline there; submit whole papers or paragraphs for rewriting and check the flagged passages; then re-test the same text range on the same platform and version.
Will it make my paper less rigorous?
Academic register, qualifiers and argument structure are kept, and terms, data, formulas, tables and references stay intact. Only filler, clichés and formulaic patterns are rewritten.
Why do CNKI and VIP give different AI rates?
They use different models, thresholds and text segmentation. In our tests, the same passage differed by up to 95 percentage points between CNKI and VIP, so rely on the platform your school specifies.
I wrote it myself. Why is the AI rate high?
Detectors judge stylistic features; formulaic lead-ins, triplets and hype are easily flagged, which does not prove AI authorship. Rewriting only targets such sentences and leaves fluent paragraphs alone.
Does it work for English papers?
Yes. Choose English as the output language; English papers follow academic-humanizer, which also covers NSF and NIH proposals, with numbers and citations unchanged.
How is this different from reducing similarity?
Similarity reduction targets text overlap with existing literature; AI-trace reduction targets the AIGC detection rate. Both can be done one after the other in the same workspace.
How much does it cost?
Sign in before starting. Rewriting runs in a workspace and deducts credits by the tokens the conversation actually uses, at 20 credits per million tokens. See the credit rules page for details.