First Research Workflow
If you have only used general chat assistants, start here to understand how NNScholar supports literature search, PDFs, study design, manuscript editing, and submission.
If you have only used general chat assistants before, start with this mental model:
A general chat assistant is mostly a question-and-answer window.
NNScholar is a research workspace. It keeps literature, PDFs, databases, manuscripts, generated files, and submission workflows inside one long-running research space.
You can still ask questions conversationally, but the result of each step should move into the right module: papers into Library, PDFs into Paper reading, manuscripts into Document Writing, submission pages into Browser, and final files into Deliverables.
The Full Path
Research question
-> Literature searches databases
-> Select papers and download PDFs
-> Library organizes a literature database
-> Paper reading opens PDFs and answers questions
-> Chat skills support study design, analysis, and writing plans
-> Document Writing drafts and revises manuscripts
-> Browser + Automatic submission assistant prepare journal submission1. Find Literature
Start in Chat by describing the research question: population, disease, method, outcome, or target field. Then open Literature.
In Literature, NNScholar searches database resources for related papers. You can review titles, abstracts, years, journals, authors, and relevance. You can also ask the agent to identify core references, priority reading, and background-only papers.
Select useful papers and save them to Library. When full text is available, use the download action so NNScholar can retrieve PDFs and organize them with the paper records.
2. Read PDFs
In Library, you can see saved papers, folders, and PDFs. Click a paper or PDF to open it in Paper reading.
Paper reading displays the PDF and lets you ask document-grounded questions:
- What is the research question?
- What are the population, sample size, and outcome measures?
- Does the method fit my study?
- Can this paper support my Introduction or Discussion?
- Which result is worth citing, and where is the page context?
The goal is not to collect loose notes. The goal is to understand and verify the paper, then bring usable evidence back into study design and writing.
3. Design The Study And Develop Ideas
Return to Chat and use the 4-phase skill strip above the input:
| Phase | What it helps with |
|---|---|
| Discovery-Ideation | Related literature, research landscape, gaps, questions, novelty, feasibility |
| Planning-Design | Study design, method selection, variables, sample size, bias and confounding checks |
| Analysis-Experiment | Data cleaning, EDA, statistical analysis, regression or causal analysis, robustness checks, figures |
| Writing-Publication | Paper structure, manuscript sections, abstract, polishing, citation checks, automatic submission |
Use these skills with saved papers, PDF evidence, and your research goal. This keeps the agent working within a research workflow instead of producing generic chat answers.
4. Draft Or Revise A Manuscript
Open Document Writing. This is not a note-taking module. It is for drafting, editing, and reviewing formal manuscripts.
Use it for:
- generating a manuscript draft from research design and evidence
- revising an existing manuscript or DOCX
- matching the format of a target journal or template
- asking questions about the document, such as whether a Methods section is complete
- reviewing logic, citations, formatting, result consistency, and overclaiming
- preparing cover letters, response letters, and submission materials
Substantive manuscript claims should trace back to papers in Library, PDF evidence in Paper reading, or reviewed analysis in Chat.
5. Prepare Submission
When the manuscript, figures, supplementary files, and cover letter are ready, return to Chat and choose:
Writing-Publication -> Automatic submission assistantThe assistant prepares the submission checklist: manuscript files, author details, funding, ethics statements, conflicts of interest, suggested reviewers, cover letter, and portal fields.
Then open Browser and go to the target journal's submission system. NNScholar can guide field-by-field review in the embedded browser, but it should pause before final submission. The final submit action should be confirmed by you.
Example Demo: Run The Full Path
Assume you are a first-time NNScholar user and want to write a manuscript about PCAT and prognosis in patients with diabetes.
- Chat: enter the research idea and ask NNScholar to split it into search, reading, design, analysis, and writing tasks.
- Literature: search PCAT, diabetes, coronary CTA, and prognosis, then identify core references.
- Library: save core papers, download PDFs, and create
core,methods, andbackgroundfolders. - Paper reading: open the first PDF and ask, "What are the population, statistical model, main results, and citable page numbers?"
- Chat skill phases: choose Planning-Design for the study design, then Analysis-Experiment for the statistical plan.
- Document Writing: create a manuscript draft, import the target journal template, run format matching, and review the document.
- Browser + Automatic submission assistant: open the journal portal and check files, author details, ethics statements, and cover letter.
- Deliverables: save the manuscript, figures, evidence table, and submission checklist.
This is the first complete research path this page is meant to help you finish.
Remember
Do not treat NNScholar as just another chatbot. Treat it as a research workspace for literature discovery, PDF downloading, literature database organization, PDF Q&A, study design, manuscript drafting, manuscript review, and submission preparation.