Everything between
“done writing” and “accepted.”
11 checks on structure, numbers and expression. Sentence-level revisions you can apply. Citations verified against real records. Submission documents, and answers to reviewers — one place for all of it.
Five things you actually do after writing.
Not a pile of features — the order you already work in. Each step ends with a file or a decision you can act on.
Find what a reviewer flags first — 11 checks.
Run them one at a time, or all at once. Findings are graded Must check / Worth checking / Improvement and each one points at the sentence it came from — so you can take it or leave it on your own judgment.
Get the revised sentence, not just the complaint.
Every finding arrives with a proposed sentence. Take the ones you want and apply them in one pass, or have them rewritten in your own writing style. Numbers, units, figure and table numbering, and citation links are never touched.
Check that every reference actually exists.
Titles, years and DOIs are matched against CrossRef, PubMed, Europe PMC and OpenAlex. If a record cannot be found it is marked as not found — a DOI is never invented to fill the gap. Retracted papers are flagged.
Meet the journal’s rules before the desk editor does.
Check the manuscript against the target journal’s requirements, run a mock review to hear what a reviewer would say, and generate the documents the submission portal asks for.
Answer the reviewers without missing a point.
Paste the reviewer comments and get a point-by-point response draft. Then check the two things that sink revisions: whether every comment was answered, and whether the manuscript really reflects what you promised.
What comes back.
Not a score — a list of what to fix, where it is and why, next to the sentence that replaces it.
Where your manuscript goes.
Unpublished manuscripts deserve a precise answer, not a slogan.
It stays in your browser
Manuscripts, checks and results are kept in this browser. Cloud storage is off by default and turns on only when you explicitly agree.
The backend does not keep it
AI requests pass through our backend to the model provider. The text is not stored or logged in our backend on the way.
Your key, your lab
You can connect your own API key. In a lab account the PI sees summary numbers only — never the text, unless you submit a version yourself.
If you turn cloud sync on, the manuscript text is stored in the project’s Supabase account, isolated per user by row-level security. Turning it off blocks upload, restore and file storage entirely.
Upload the manuscript.
See what a reviewer would see.
Review, revise, verify, submit, resubmit — in one place, in the order you already work. Nothing is applied until you say so.
Go to the app