An AI that knows academic writing
It knows what an SCI paper, a thesis and a grant proposal should each look like. Find papers, write the first draft, revise with your team, all the way to submission, with every citation opening the original.
Thanked in papers published inTheranosticsFrontiers in NeurosciencePoultry ScienceAging MedicineInteractive Journal of Medical ResearchKnows what a paper should look like
Each document type and each kind of review has its own writing skill.
- Picks one of four scientific question attributes: original exploration; frontier with a distinct approach; demand-driven bottleneck breakthrough; cross-disciplinary common problems. Most Young Scientists and General projects fit the second
- Builds the rationale in five layers: background 1–2 paragraphs, field review 3–5, open problems 1–2, key scientific question 1, this project's entry point 1; cites representative work from the last 2–3 years
- Writes 2–4 innovations as "to address…, this project will…, thereby…" and checks each for vagueness, technique dressed up as innovation, overreach such as "filling a domestic and international gap", or drift from the aims
- Background1–2 paragraphs: why the field matters
- Field review3–5 paragraphs: prior work, grouped by sub-direction
- Open problems1–2 paragraphs: what is unsolved and why it is hard
- Key scientific question1 paragraph: the core contradiction
- Entry point1 paragraph: how this project tackles it and why it can
- Builds the abstract from Nature-style sentence jobs: field context 1–2 sentences, detailed background 2–3, problem 1, main result 1, implications 2–3, broader context 1–2, about 190 words
- Keeps titles within 75 characters and figure legends within 300 words, and defines abbreviations on first use
- Flags overclaiming words such as prove, unprecedented, first, best and superior, and rewrites them to what the evidence supports
- Field context1–2 sentences: readable outside the subfield
- Detailed background2–3 sentences: introduces subfield terms
- Problem1 sentence: unmistakable
- Main result1 sentence: the main claim, not the method pipeline
- Implications2–3 sentences: what the result changes
- Broader context1–2 sentences: back to the wider question
- Deletes stock transitions such as "in summary", "it is worth noting that" and "furthermore", letting each sentence follow from the last
- Rewrites colon lists ("first…; second…; third…") and colon tails ("this raises a key contradiction: …") as natural prose
- Replaces unquantified modifiers such as "significantly improves" with numbers or deletes them; section titles drop "Experiment 1:" prefixes and "A and B and C" splicing
The network has deep feature extraction capability, significantly improves recognition accuracy and lays a solid foundation for subsequent tasks.
The network reaches 92.3% recognition accuracy on the test set, 3.1 percentage points above the baseline.
- Scores NSFC proposals out of 100 on four dimensions: rationale and content 40, plan and feasibility 30, innovation and value 20, foundation and conditions 10; advice is graded critical, important or suggested
- Checks 10 common deductions, such as whether aims match content, innovations map one-to-one to content, recent 2–3 years of literature are cited, and the budget matches the plan
- Simulates three reviewers at Nature standards for SCI manuscripts and checks that the same number matches across text, tables and legends
- Rationale and content31 / 40
- Plan and feasibility22 / 30
- Innovation and value14 / 20
- Foundation and conditions7 / 10
[Critical] Innovation 3 has no matching research content; merge it into innovation 2 or add the content
These methods come from 23 academic writing skills; the two most popular open-source projects behind them each have over 50,000 GitHub stars. See all skills
References are searched from scholarly databases
Search covers journal articles, arXiv preprints, PubMed and Chinese literature. Chinese journals are tagged CSCD, PKU Core and CSSCI; English journals show their CAS quartile and impact factor.
Citations open the original
Paper citations link to the DOI; citations of your uploaded PDFs jump to the exact page.
Import your existing library
Import from Zotero and EndNote, or upload PDFs and Word files directly.
Exported Word keeps citations updatable
The reference list is built with the text, and exported Word files carry citation fields for Zotero or EndNote.
AI drafts and revises with you
Say what to write and which comments to address. The AI writes into the document and fixes the text, several edits in one go.
Revise with your supervisor and classmates
Send one invite link and your supervisor and classmates join the same draft, seeing who is reading and who is editing.
One draft, edited together
Everyone's cursor and name appear in the text and changes sync live, with no files sent back and forth.
Hand comments to the AI
Your supervisor comments on selected text; tick a few comments and choose Hand to AI to revise accordingly.
From proposal to defense, in one place
Many SCI papers thank Acadwrite
- 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.
Comparisons with other tools
FAQ
Does Acadwrite make up references?
Papers are searched from scholarly databases. Every citation in the text links to the original paper or a page of your uploaded PDF so you can check it. When adding citations to paragraphs, it checks the original and marks weak support as pending instead of citing it.
How is it different from writing papers with ChatGPT or DeepSeek?
General chat tools search the web. Acadwrite searches over 300 million academic papers, applies the conventions of each document type, and keeps papers, drafts and chat in one workspace.
What can it write?
SCI papers, theses, introductions, NSFC, NSF and NIH proposals, research proposals, responses to reviewers, conference papers and defense slides each have their own method; you can also just say what you need in the chat.
How do I revise with my supervisor?
Click Invite at the top of the workspace and send the link. Everyone edits the same draft at once, selects text to comment, and can tick comments to hand them to the AI.
Can I submit the draft as is?
Your core arguments and conclusions should be your own. Before submitting, read the full text, check the citations and disclose AI use as your university or journal requires.
How much does it cost?
You pay by actual use at ¥2 per credit, with free credits for new accounts. A paper review costs 8 credits, adding citations to a passage 5, and polishing a document 1.