How to Prepare a Manuscript Faster (Without Using AI)

Preparing a manuscript involves a surprising amount of admin. Author lists, affiliations, disclosures, funding statements, and references all need to be added before publication.

This post covers the 8 free tools I use for those tasks; most are built by researchers, and none use AI.

If you’re first author on a multi-author paper and the admin *yawn* is starting to look intimidating, this article is for you.

The Best Tools for Manuscript Admin Are Not AI

Let me address the elephant in the room: I prefer not to use AI for these tasks.

My reasoning is that whatever an AI tool produces for me, I still have to check that it is right. For something like an author list, checking the work of a large language model (LLM) is a massive task! So, for me, the time saved by using AI is less than it looks.

The purpose-built tools I am about to cover were written by researchers for a specific job, and if they get something wrong, they do so within the boundaries of what you originally supplied. Unlike LLMs, they won’t hallucinate a citation or risk inventing affiliations that don’t exist.

Automating VS Delegating

To understand the advantage of delegating your work rather than automating it, let’s look at the NIH AuthorArranger.

This tool doesn’t write your author list; instead, it gives you a blank table. You send that table to your co-authors to fill in their own affiliations, and the tool assembles what they gave you into a formatted list. This tool isn’t producing anything you couldn’t produce yourself; it’s just delegating the work for you.

So the test I now apply to any tool that is part of my manuscript process isn’t how much of this can it do for me? It’s: does this make the work delegable?

The Five Stages of Manuscript Preparation

Here’s how I think about the whole arc of the manuscript writing process (table 1). The first column shows the five stages, the middle column is what to do while you’re there, and the right-hand column is what you need to know before you move on to the next stage of the process:

Stage	Do this while you’re here	Settle this before you leave
1. Review the literature	Read for what’s already been done, so your experiments target what’s genuinely new. Meet co-authors regularly (monthly at minimum) to get a steer early rather than a correction late.	Do you have a target journal in mind? Does one journal keep appearing in your citations?
2. Research	Keep a detailed record of reagents: catalog number, lot number, method. Write the methods section as you go. Outline your figures — they’re the backbone of the paper.	Where does the raw data for each figure live, and is that written down anywhere but your memory?
3. Write	Draft the manuscript, figures and legends. Make a graphical abstract to give editors, reviewers and a wider audience a visual summary. Set interim deadlines for individual sections.	Does your institute cover article processing charges, and what are your funder’s open access requirements?
4. Revise	Send the full draft to co-authors with a deadline — two weeks is reasonable.	In that same email: request conflicts of interest, journal contact disclosures, funding acknowledgments, and the Author Arranger table.
5. Submit	Upload to a preprint archive if you’re using one, following the archive’s own submission requirements and checking your target journal’s preprint policy. Set aside a whole day for the journal portal.	Has someone outside the project proofread everything, including the text you typed into the portal fields?

1. Review the Literature

“A month in the laboratory can often save an hour in the library.” 

Given the choice between reading and pipetting, most researchers will default to pipetting. It’s worth noticing when you’re doing that so you can stay focused on the literature.

2. Research

“There are two kinds of people: those who back up their data, and those who will.”

I learned this one the hard way before I had a result archive I could navigate, i.e., a master index telling me which raw files produced which figure. From experience, that’s the question you actually get asked, sometimes years later, and a folder of safely-backed-up data doesn’t answer it…

3. Write

“Work expands so as to fill the time available for its completion.”

During this process, I do think it’s very useful to set mini-deadlines to keep you on track, so you can set dates for yourself for when you’ll have different sections done by specific times. If you’re staring at a blank document and not feeling inspired to write, start with the methods section. It’s the section where nobody wants beautiful prose, so there’s no pressure for things to be perfect. The bar is so low that you can’t fail at it, making it the easiest possible way to get words on the page.

4. Revise

“One email is worth four chases.”

Your co-authors’ attention is finite, and every single thing you ask them for is a separate ask for someone who has their own deadlines to attend to.

The first time you circulate the full draft of your manuscript is the most valuable email you’ll send. So, along with the request for feedback, ask for everything else you’ll need: conflicts of interest, any journal contact disclosures, and all funding acknowledgments, collected in one place. It costs almost nothing to add to an email I’m sending anyway, and it saves four separate chases in six weeks’ time!

5. Submit

“A paper in print is worth two in development.“

It’s really easy for many things in academia to create competing deadlines, but getting your papers published is crucial, and being able to show that you can complete these projects and share the information with the world is extremely important. Two brilliant manuscripts on your hard drive are worth less than one merely good paper that’s out. The unpublished one doesn’t count on a CV, and nobody can read it.

Another consideration is article processing charges. This is something I genuinely didn’t know about early in my career. If that is you, keep in mind that some open-access routes charge a publication fee, and some don’t. Many institutes and funders have funds or agreements to cover those costs, and that conversation with your open-access team usually needs to happen before you submit, rather than after an acceptance email arrives with a massive bill attached!

Tool 1: NIH AuthorArranger

If you’ve ever built an author affiliation list by hand, you’ll know it’s fiddly and slow. To help, you can access the NIH AuthorArranger here. Download the template, share it with your co-authors, and have each of them enter their own name and affiliations, including those with multiple affiliations. Paste the completed table back in, and it produces the author list with superscripts mapped to the affiliation list, in the order you specified. This tool is also useful for conference posters.

Tool 2: Reference Managers

Reference managers such as Zotero, Mendeley, JabRef, and Lean Library Workspace (formerly SciWheel) will all handle a library, generate a bibliography in a very large range of journal and citation styles, and insert in-text citations. Most run on a freemium model (free to use, with paid tiers for things like extra storage), so “free” is broadly true, but worth checking against your own needs before you commit a whole project to one.

I’m most familiar with Zotero, as it does everything I need. There’s a Word plugin for in-text citations, the library syncs between the browser and your desktop, and libraries can be shared with co-authors. That last feature works on the same principle as the AuthorArranger: it’s making one library the single source of truth for everybody.

For the occasional awkward source (e.g., a website, a report, something with no DOI), the Free Harvard Referencing Generator from MyBib is a quick way to get a usable citation you can drop into your library.

Tool 3: anystyle.io

If you have a document assembled by several co-authors, the references are plain text with no links, and you need a single, coherent bibliography in any style, AnyStyle.io (pun intended) is the tool for you.

Paste your raw bibliography text into AnyStyle, and it will parse it into a structured file (e.g., BibTeX) that you can import into your reference manager. With something on the order of a hundred citations, I’d estimate it has saved me at least ten hours compared with searching for and adding each one by hand! It works from what you paste in, so the output is worth a check before you publish it.

An Important Caveat

It’s worth keeping in mind that AnyStyle can’t restore links between your in-text citations and the library entries. The software can rebuild the bibliography itself, but when a document arrives as plain text with those reference-manager links already lost, it can’t restore them. Once those structured citation links have been stripped out, there is nothing in the plain-text bibliography that tells AnyStyle how each in-text citation was linked to the library.

If you can agree with your co-authors, at the start, on one free reference manager that everyone uses, you never enter this situation at all. It’s a short conversation, but it has the highest return of anything else in this article!

Tool 4: JANE

JANE (Journal Author Name Estimator) can help when you are deciding where to publish your manuscript. It takes your abstract and returns a ranked list of journals where similar work appears. JANE ranks journals by a confidence score based on how closely your text matches articles already in PubMed.

There are two key things I like about it:

1. If you’re worried about pasting an unpublished abstract into a web tool, there’s a scramble option that reorders the words into alphabetical order in your browser. This keeps the words for matching but makes the original text hard to read, protecting your work.
2. It surfaces journals that weren’t already on your radar, which is useful if you’re new to a field, as you can work through their tables of contents to get oriented.

When submitting your paper, I think it’s worth aiming a little high initially. You can submit to an ambitious journal first and work down if it doesn’t get accepted. Just weigh that against fit and audience, and against the time each round of rejection costs. The prestige isn’t worth chasing if your paper clearly belongs somewhere else!

Tool 5: NIH BioArt

NIH BioArt is a free library of publication-ready scientific and medical illustration icons. It includes cells, vessels, organs, and lab apparatus. It’s free to use in publications, though some illustrations require specific credit, so check the requirements for each item.

A graphical abstract summarising your key findings can help bring in a broader readership, and it’s a courtesy to editors and reviewers who are trying to see quickly what you found. If you want to do it well, “Ten Simple Rules for Designing Graphical Abstracts” (Jambor & Bornhäuser, 2024) is worth reading first.

Tool 6: G*Power

G*Power is free software for power and sample-size calculations across a range of statistical tests and experimental designs. Broadly, you supply an expected effect size, and it returns the sample size you’d need to detect it at your chosen power and significance level.

You can inform that expected effect size with prior evidence or pilot data. But treat a pilot estimate with caution, because effect sizes from small pilot studies are unstable and shouldn’t be plugged in as though they were the true effect. Power is commonly set at 0.80 (sometimes 0.85) and alpha at 0.05, but those are conventions rather than rules.

Doing this at the start of your work lets you say later that the study was designed to answer a specific statistical question, instead of trying to retrofit the data later.

Tool 7: The Open Science Framework

The Center for Open Science runs the Open Science Framework, where you can preregister your study. To register your work, you’ll need to include your question, hypotheses, and the statistical plan you intend to follow.

I haven’t done this myself, but several of my colleagues have. I mention it partly because it’s a good way to get moving when you don’t know where to start, since it forces you to write down the scientific method of what you’re actually doing!

Tool 8: Your Field’s Own Checklists

One of the most useful things about science is that researchers have already written the guidelines…and they’re free! The EQUATOR Network (Enhancing the QUAlity and Transparency Of health Research) collects reporting guidelines by study type. The four I have listed below are the ones I come across most, but there are many more. Find the one specific to your paper and work to it from the start, rather than retrofitting at the end.

Guideline	Full name	What it’s for
ARRIVE	Animal Research: Reporting of In Vivo Experiments	Reporting research involving animals (in vivo studies)
PRISMA	Preferred Reporting Items for Systematic Reviews and Meta-Analyses	Systematic reviews and meta-analyses (scoping reviews use the PRISMA-ScR extension)
CONSORT	Consolidated Standards of Reporting Trials	Randomized controlled trials
CARE	CAse REport guidelines	Clinical case reports

Where I Use AI in Manuscript Preparation

AI is great for tightening a paragraph you wrote or suggesting a clearer way to phrase a sentence. In those cases, there’s no external source of truth to reconcile against, so the checking is just reading (which you were going to do anyway!)

What I’d avoid using AI for is anything about who your co-authors are, what they’re affiliated with, who funded them, what a specific journal requires, or which of your figures came from which experiment. Using AI to retrieve this information removes important guardrails and increases the risk of hallucinations.

Where to Put the Effort

If you take one thing from this, I’d like it to be this: there are only two real levers for manuscript preparation:

1. The first is sequencing: do each task at the point where the information is still fresh, e.g., write the methods while the reagents are in front of you, index the raw data while you remember where it came from, and ask about article processing charges before submitting your manuscript.
2. The second is delegation: for anything that lives in someone else’s head, your job isn’t to find it out but to make it easy for them to tell you. That’s the blank table, the shared library, the single well-loaded email.

The tools here are time-savers that help make the process smoother. Good luck with the writing and with the submission portal. And make sure you set aside a whole day for this admin; you’ll need it!