Jenni AI vs QuillBot: Which Revises a Thesis Paragraph Better?

You have a chapter drafted and need to know whether its argument holds before your supervisor reads it.

Jenni ran 2 reviews on the test paragraph in 76 seconds, and Peer Review listed the double hedge 'perhaps maybe' among its 4 weaknesses.

With no review of that kind in its menu, QuillBot's part was the wording, where the Fluency rewrite fixed the agreement error in the clunky sentence.

<CTA title="Get a peer review of the chapter before you send it" description="Peer Review scores the draft, lists its weaknesses and pins comments to the sentences they concern" buttonLabel="Try Jenni Free" link="https://app.jenni.ai/register" />

TLDR: Jenni or QuillBot?

Draft, cite and check in Jenni. It filed the trial into the sentence and the reference list in one action, ran 2 reviews in 76 seconds and exports to Word or LaTeX. Use QuillBot when the file must stay in Word, through its add-in, or to build a reference from a title on a page of its own.

How the two were tested

Both tools worked on the 201-word paragraph on lipid nanoparticle delivery of CRISPR-Cas9 used across these tool reviews, on the same day.

• 2 planted faults - the sentence "There is a lot of things that still needs to be worked out though" and the double hedge "could perhaps maybe".

• 1 claim that needs its source - the mean 87% fall in serum transthyretin at the higher dose, from the 2021 NEJM report of the NTLA-2001 trial.

• What each tool was given - Jenni had the whole paragraph in one document. QuillBot's Paraphraser got the 3 sentences around the faults, 96 words, its Citation Generator got the trial's title, and AI Chat got the trial sentence.

• What the screen showed - quotes were copied from the screen and word counts were taken from the copied text. Neither Jenni's panels nor QuillBot's AI Chat showed a model name.

Jenni took the 4 tasks in the order below, each on the draft as the task before had left it.

When a sentence reads badly

Jenni

The planted sentence runs to 43 words: "There is a lot of things that still needs to be worked out though, for example how the particles behave outside the liver, whether repeat dosing causes an immune response, and what the long term editing outcomes look like at off target sites."

Selected and sent to AI Edit with the Improve fluency command, it came back in 4 seconds as 28 words:

"Several questions remain, including how the particles behave outside the liver, whether repeat dosing triggers an immune response, and what the long-term editing outcomes are at off-target sites."

The Jenni rewrite, with deletions crossed out in red and additions in green

See edits marked each deletion and each insertion, and a What changed and why note explained the swap. Replace Selection put the new sentence into the draft, and the word counter went from 207 to 192.

QuillBot

The Paraphraser in Fluency mode took the trial sentence, the clunky sentence and the hedge together, 96 words in and 83 out. The first output appeared at 9.4 seconds and settled by 12.3, and its version of the clunky sentence runs to 35 words.

"However, many questions remain to be answered, such as how the particles behave outside of the liver, whether repeat dosing causes an immune response, and what the long-term editing outcomes look like at off-target sites."

The agreement error and both missing hyphens are fixed. The same pass also removed the planted double hedge, giving "It is possible that ionizable lipid chemistry is the primary mechanism for extrahepatic targeting, but this has yet to be demonstrated conclusively." A second Fluency run returned the same text word for word.

Winner - Jenni, for the tighter sentence with its change explained. Improve fluency cut the 43 words to 28 and wrote out what it changed, where the Fluency version of the same sentence ran to 35.

<ProTip title="🔤 Check every swapped word:" description="Fluency swapped lever for mechanism in the hedge. A single noun can change what a sentence claims, so read each swap against the paper behind it" />

The guide to paraphrasing in academic writing covers how far a rewrite can go before the meaning moves.

When the claim needs its source

Jenni

With the cursor after "after a single infusion.", the Cite button opened a search at the cursor with All, Discover and Library tabs. Before anything was typed, Discover listed 20 papers drawn from the sentence, with the NEJM trial 20th.

Typing the trial's title brought back 18 result cards. The first was a 2022 reprint in the Yearbook of Pediatric Endocrinology with the same title, and the NEJM report came second, showing 1.6K citations and an impact factor of 24.39. Each card carried Cite, View and Save buttons.

One click on Cite in the NEJM card put "(Gillmore et al., 2021)" into the text, just after the sentence's full stop, and started a References section under the draft set to APA (7th ed.).

The in-text citation in the paragraph, and the reference it created under the draft

The entry lists the first 19 authors, an ellipsis and the last author, the APA 7 form for 21 or more authors, and ends "385(6), 493–502. https://doi.org/10.1056/nejmoa2107454", matching the PubMed record.

QuillBot

The Citation Generator takes a title, URL, DOI, ISBN or keywords and offers more than 1,000 styles, with APA 7th edition as the default. The same title search returned 20 results, with the NEJM report first and the 2022 reprint second.

Each row showed the title, the type, the year and the author surnames. No row showed a journal name, a citation count or an impact factor, so the reprint and the trial were told apart by the year.

Choosing the first result opened a References page with the entry, a Suggestions tab reading "All good!", and the parenthetical and narrative in-text forms beside it, each with a Copy button.

The QuillBot reference list, on its own page

The entry follows the same 19-author form that Purdue OWL sets out, with volume, issue, pages and DOI. It prints the title as "CRISPR-CAS9 in vivo gene editing" and the third author as Taubel, where PubMed has Täubel.

From the claim alone, QuillBot's AI Chat with web search, asked for the source and its APA 7th reference, answered in 19 seconds with the June 2021 Intellia and Regeneron press release. Its figure was right, 87% at 0.3 mg per kilogram, and it gave no DOI.

<ProTip title="📰 Trace a press release to its paper:" description="A company announcement can carry the right figure and still be the wrong citation. Find the journal article it reports, here the 2021 NEJM trial, and cite that" />

Winner - Jenni, for a citation and its reference placed in the draft in one action. One click filed the in-text citation and the full reference with its DOI under the paragraph, while QuillBot built the entry on a separate page for copying across.

<ProTip title="🔗 Match the record to the DOI:" description="Both searches returned the 2021 trial and a 2022 reprint with the same title. The trial is the one whose DOI ends NEJMoa2107454" />

The guide to primary sources covers why the paper, not the announcement, is the citation a thesis needs. A paper already in a reference manager can come in by DOI or through the Zotero and Mendeley import.

When the argument has to hold up

Jenni

Claim confidence ran on the draft as it stood, with the sentence fixed and the trial cited. It searched academic databases with 3 queries of its own and finished at "Thought for 28 seconds" with 2 suggestions, both shown in the text before anything was accepted. Its breakdown marked 4 other categories clear: Misrepresented, Contradicted, Overstated and Unverifiable.

• Unsupported - the apolipoprotein E sentence had no citation, and the review proposed Hosseini-Kharat et al. (2025), a review article in Molecular Therapy Methods and Clinical Development on why lipid nanoparticles target the liver. Accepted.

• Weakly supported - it proposed swapping the NEJM citation for a 2022 article that reports the trial second hand, saying the cited excerpt lacked the 87% figure. The figure is in the trial abstract, so this one was rejected.

The accepted source went into the reference list in one click, with its DOI.

The source Claim confidence proposed, filed under References with its DOI

Peer Review then ran in 48 seconds and scored the paragraph 7/10 overall, with Soundness 2/4, Presentation 3/4 and Contribution 3/4. It listed 3 strengths, 3 questions for the authors and 4 weaknesses, one of which named the planted fault.

"Uses informal and overly hedged language, such as 'perhaps maybe,' which detracts from the scientific rigor and authoritative tone expected in gene therapy literature."

Its fourth weakness, and 1 of the 2 comments it pinned to sentences, repeated the Claim confidence reading that the cited excerpt lacked the 87% figure.

QuillBot

QuillBot's menu includes the Paraphraser, Grammar Checker, AI Chat, Translate, Summarizer, Citation Generator and Quillbot Flow. None of them is a review that scores a draft's argument or checks its claims one by one against sources.

Winner - Jenni, for a review of the argument that lands in the draft. Claim confidence and Peer Review ran in 76 seconds between them, filed a source with its DOI and pinned their comments to the sentences concerned.

<ProTip title="🧾 Read the entry against PubMed:" description="The QuillBot entry printed CRISPR-CAS9 and Taubel. Open the PubMed record and compare names, capitals and page numbers before the reference list is final" />

When the chapter goes to your supervisor

Jenni

The three-dot menu holds Export, labelled Word or LaTeX. The dialog offers Word (.docx) and LaTeX (.tex), and the Word tab offers MS Word native citations or Hyperlink citations.

Both tabs carry the same layout presets: Jenni default, Double-spaced manuscript, Two-column paper, and Thesis or report, which in Word adds a title page and a table of contents. There is no PDF option in the dialog.

QuillBot

The document editor that the Grammar Checker opens in had a copy icon in its toolbar. Its three-dot menu held Preferences, Clear text, Give feedback and Move to trash, with no export or download item.

The reference list page has its own Export button for Word, which covers the references only, and the Summarizer carries an Export icon for its output.

Winner - Jenni, for the choices at the export step. The dialog offered Word with native citation fields or hyperlinks, LaTeX, and 4 layout presets including a thesis layout, against a copy icon in the QuillBot editor.

Where Jenni is ahead

• The citation lands in the draft - the in-text citation and the full reference were filed together from one result, with the DOI attached.

• The argument gets reviewed - Claim confidence and Peer Review ran inside the document in 76 seconds between them, with scores, a filed source and comments pinned to sentences.

• The file leaves as Word or LaTeX - with native Word citations and a thesis layout preset, from the same place the chapter was written.

• The edit came with its reason - AI Edit showed what it removed, what it added and why, before anything changed in the draft.

The reviews run on the open draft, so a check like this needs no copying between tabs.

What QuillBot has that Jenni does not

• A Word add-in - QuillBot lists one among its integrations on its pricing page, so it can be used inside Microsoft Word. Jenni does not offer an add-in; it imports .docx files and exports Word documents with native citations instead.

Everything else QuillBot showed in this test, from rewording to reference entries, has a counterpart in the AI Edit commands and Cite.

What each tool costs

Prices from jenni.ai/pricing, quillbot.com/premium and the QuillBot pricing FAQ.

Jenni Plus adds 500 AI Edits and 10 Reviews a month, and Pro makes edits, chat and Reviews unlimited, both with full document export. QuillBot Premium buys full access to the Paraphraser with 9 modes, summaries up to 6,000 words and higher AI Chat usage, and QuillBot advertises its Citation Generator as free and without limits.

Which one should you use?

The PhD student halfway through a thesis

Write in Jenni. A thesis chapter carries dozens of claims, and here a citation went into the text and the reference list in one action, then 2 reviews checked the argument before the chapter left the editor.

The researcher with a manuscript for a journal

Draft and check in Jenni, using Peer Review for the argument and Claim confidence for the sources. If the file must stay in Word, QuillBot has a Word add-in; if you want one workspace that drafts, cites, reviews and exports, write in Jenni.

The Masters student whose supervisor asks for every source

Jenni, for the citation in the sentence and the reference underneath. Jenni is built for the thesis and the manuscript; it will handle an essay, but it is the citation and review workflow a dissertation needs.

The Jenni vs Paperpal comparison covers a tool that works inside Word, and the guide to proofreading a thesis covers the last pass before submission.

<CTA title="Export the chapter as Word or LaTeX" description="Choose native Word citations or a LaTeX file, with a thesis layout preset that adds a table of contents" buttonLabel="Start Writing Free" link="https://app.jenni.ai/register" />

For the chapters still to come, start the next one in Jenni, where the draft, its sources and its reviews stay in the same file from the first sentence to the export.