ARCHITECTURE / LITERATURE
From question to traceable review
- 01
Question & criteria
Scope, terms, inclusion and exclusion
- 02
Source discovery
Web, databases, citation exploration
- 03
Exported records
Retain sources and search details
- 04
Record processing
Exact deduplication, year filter, sort
- 05
Staged screening
Title → abstract → full text
- 06
Extract & assess
Study characteristics, findings, quality
- 07
Thematic synthesis
Agreement, differences, limits, gaps
- 08
DOI metadata check
Identifier lookup and metadata report
- 09
Compare & revise
Compare citations with original sources
- 10
Review document
Markdown / PDF with evidence records
01 / LITERATURE-REVIEW
Start with a question and an evidence plan
Literature Review organizes the preparation of a research synthesis around a defined question. The process begins with the review type, time window, core concepts, synonyms, and explicit inclusion and exclusion criteria. In the AutoResearch catalog, this module occupies the evidence-preparation stage: mapping existing methods, findings, and disagreements before a new direction is developed. The underlying toolkit provides a structured procedure, reference guidance, a reporting template, and focused scripts for recurring tasks. A clear scope also keeps discovery, selection, and writing connected to the same question throughout the review.
02 / LITERATURE-REVIEW
Discover sources and prepare the record set
Discovery combines broad academic web search with complementary databases, preprint sources, and forward or backward citation exploration. The selected sources depend on the topic; the workflow asks for search dates, query strings, result counts, and retained source records. Exported results then become the input to the local processing utility. It can remove exact duplicates using DOI or title rules, apply a year range, sort supplied citation, year, or relevance fields, and produce JSON, Markdown, or BibTeX. Database discovery itself is supplied by external tools. Pilot searches and revised terms help calibrate the source set before a broader search is carried out and documented.
03 / LITERATURE-REVIEW
Screen studies and extract comparable evidence
Selection proceeds through titles, abstracts, and full texts against the agreed criteria. Exclusion reasons and retained-study counts belong in the review record. For included studies, the reporting structure captures design, population or sample, methods, findings, author-reported limitations, funding, and conflicts of interest. It also provides places to document a study-appropriate quality assessment. These steps require research judgment and access to the relevant texts. The toolkit supplies the procedure and fields; completed assessments, reviewer decisions, and their supporting records are created during the review. A common extraction form helps make studies comparable while retaining the details that explain meaningful differences between them.
04 / LITERATURE-REVIEW
Build a synthesis across studies
The writing stage connects evidence across studies through themes or research questions. It compares agreement and disagreement, explains differences in methods, and identifies open questions with their supporting sources. Study tables and extraction records help readers move from a synthesis back to individual evidence. The template includes sections for methods, findings, discussion, limitations, and appendices. Search-flow or conceptual diagrams can make the structure easier to inspect. Optional meta-analysis fields are available for reviews that separately perform the necessary statistical work. The strength and scope of each conclusion should stay aligned with the evidence actually included in the review.
05 / LITERATURE-REVIEW
Review references and produce the document
Before delivery, the DOI utility checks identifiers and retrieves Crossref metadata into a report. That report supports a separate comparison of authors, titles, dates, and citation use with the manuscript and original sources. The completed review remains editable in Markdown and can be exported to PDF when the document tools and bibliography assets are available. Final inspection covers references, figures, tables, and layout. The result is a review package with a visible evidence trail. Systematic-review reporting requirements and substantive claim checks must be completed and documented for the particular project. When an error is found, the workflow returns to correction and rechecking before the affected references are finalized.
Outputs & scope
- Search and selection records
- Study tables and extracted evidence
- Thematic synthesis and open questions
- DOI metadata review report
- Markdown and PDF documents
This page describes a workflow and its supporting utilities. Source access, selection decisions, and substantive evidence checks remain part of each review.
Underlying toolkit: K-Dense · MIT license. Presented here as an evidence-preparation module in the AutoResearch catalogue.
These modules describe the programme’s intended division of work. Cross-module runtime connections are still being integrated and validated.
Discuss the project