Where do I paste my manuscript, and how much should I include?

Paste your manuscript directly into the Manuscript Content field on this page — no longer in the chat. At minimum include the complete Methods and Results sections. Adding the first paragraph of Discussion (the main conclusion) helps the AI write a more accurate Conclusions section. Do NOT paste the full paper: the template is optimized for focused section input (soft cap ~2000 words, hard cap 4000). Providing everything up-front lets AI Run generate the abstract in a single step without wasting a credit on back-and-forth.

What does [DATA NEEDED] mean in the generated abstract?

It flags information that should be in the abstract but was not found in your input. Common triggers: missing sample size, missing primary outcome results, missing follow-up duration. Fill these in from your paper before submitting.

Which format should I choose for a retrospective cohort study?

Use STROBE format — it is the international reporting standard for observational studies (cohort, case-control, cross-sectional). In TARGET_JOURNAL_FORMAT, write: "STROBE observational cohort, [word limit] words."

My journal wants an unstructured abstract. Can this generate that too?

Yes. In TARGET_JOURNAL_FORMAT, write "Unstructured, [word limit] words." The output will be a single flowing paragraph without subheadings, maintaining all content rules (no overclaiming, data fidelity, word count guard).

Can I use this for a systematic review or meta-analysis?

What is "spin" in medical literature, and how does this prompt handle it?

"Spin" is the use of language to overstate positive findings or downplay negatives — e.g., leading with a secondary outcome after the primary endpoint failed, or labeling a non-significant trend "promising." Studies show LLMs often amplify spin rather than correct it. This prompt counters this with Rule 5 (Spin Guard): before drafting, the AI explicitly scans your manuscript for spin patterns and reports them in the Pre-Generation Analysis. The abstract is anchored only to statistically significant, pre-specified outcomes.

How does the generated abstract connect to my full journal submission workflow?

The finalized abstract feeds directly into downstream tasks: (1) paste it into the Cover Letter Generator to introduce your study's key findings; (2) use the suggested MeSH keywords for PubMed/Embase submission tags; (3) if rejected, the structured abstract gives a rebuttal precise context about your study design and results. Tip: save the finalized abstract alongside your manuscript version so all submission materials stay in sync.

Do I need to disclose AI assistance when submitting the finished work?

Yes — ICMJE (2023) guidelines and most major journals (NEJM, Lancet, Nature, JAMA, BMJ) require authors to disclose AI-assisted writing in the Acknowledgements or Methods section. AI tools cannot be listed as authors. A typical disclosure: "AI-assisted writing tools [specify name] were used in the preparation of [specify sections] of this manuscript. The authors reviewed, revised, and take full responsibility for all content." Check your target journal's specific policy, as requirements vary.

The AI output uses Markdown (`**bold**`, `##`, lists) and Unicode superscripts (`²`, `₂`). How do I convert it to journal-ready Word format?

Pasting Markdown directly into Word leaves literal `**` and `##` characters, and Unicode superscripts/subscripts are NOT equivalent to Word's native superscript formatting — they render in a different font and cannot be batch-reformatted by the journal's typesetter. Three workflows (Windows | macOS): **1. Pandoc — fastest and cleanest.** Install: `winget install pandoc` (Windows) or `brew install pandoc` (macOS). Convert: `pandoc input.md -o output.docx`. Handles headings, bold/italic, bullets, and tables → native Word styles. Unicode sup/sub still need a manual pass (see step 3). **2. Online Markdown → DOCX converter.** Paste into any "Markdown to DOCX" web tool, export the `.docx`. Works identically on both platforms. No install needed. **3. Manual cleanup inside Word.** - Find & Replace (remove `**`, `##`, `- `): Windows `Ctrl+H` | macOS `⌃H` or `⌘⇧H` - Bold / Italic: Windows `Ctrl+B` / `Ctrl+I` | macOS `⌘B` / `⌘I` - Superscript (for `x^2` → x²): Windows `Ctrl+Shift+=` | macOS `⌘⇧=` - Subscript (for `x_2` → x₂): Windows `Ctrl+=` | macOS `⌘=` Handling Unicode `²` `³` `₁` `₂`: Find & Replace them with the plain digit, select the digit, apply the superscript/subscript shortcut above. For prompts that output ASCII notation `x^n` / `x_n` (e.g., abstract-generator v1.4+), follow the table directly.

Structured Abstract Generator -- medPrompt

Usage Guide

Fill in study type, target journal format, and word limit.
Paste your manuscript (at minimum Methods and Results) into the Manuscript Content field.
Click AI Run — the assistant generates the structured abstract directly, no chat back-and-forth needed.
Resolve any [DATA NEEDED] placeholders before submission, and review the post-abstract checklist.

Medical Research Assistant

Fill in variables and run directly with AI

Wiki

The abstract is the single most read section of any published paper — for many readers it is the only section they will ever see. A poorly written abstract that oversimplifies findings, inflates conclusions, or mismatches its target journal's format can lead to desk rejection or distort how the work is cited and interpreted. This tool generates precise, structured abstracts for medical manuscripts with built-in quality rules that prevent the most common abstract errors.

The tool supports four major abstract formats: IMRAD (for general medical journals), CONSORT (for randomized controlled trials), STROBE (for observational cohort and case-control studies), and PRISMA (for systematic reviews and meta-analyses). It also generates unstructured single-paragraph abstracts for journals that require them. For each format, the correct section labels and content requirements are applied automatically.

A core design principle is data fidelity. The AI will not hallucinate statistics or fabricate effect sizes. If a required data point (sample size, primary outcome result, follow-up duration) is missing from your pasted manuscript, it inserts a [DATA NEEDED] placeholder rather than inventing a number. These placeholders serve as a checklist: resolve each one before submitting to a journal.

The Spin Guard rule performs a pre-generation analysis of your manuscript for spin patterns before the abstract is drafted. This includes detecting conclusions anchored to non-significant results, secondary outcomes promoted above the primary endpoint, and vague "promising" language around null findings. The abstract is anchored only to statistically significant, pre-specified outcomes.

After the abstract is generated, the tool provides a post-abstract quality checklist covering word count compliance, structure completeness, and keyword suggestions. Five MeSH-compatible keywords are suggested automatically to support PubMed and Embase indexing.

This tool works best when the complete Methods and Results sections are pasted as input. The first paragraph of Discussion may also be included to improve the Conclusions section. Do not paste the full paper — the template is optimized for focused section input.

FAQ