How to Import PDF Collections and Configure Concept Tags in Lateral for Systematic Literature Screening

Academic researchers and graduate students conducting systematic literature reviews frequently encounter a structural bottleneck: traditional reference managers excel at organizing bibliographic citations, but they offer little help when extracting and synthesizing specific evidence across dozens or hundreds of PDF files. Researchers are usually forced to open papers one by one, run repetitive Ctrl+F keyword queries, and manually copy-paste quotes into messy spreadsheets.

Lateral was engineered specifically to solve this cross-paper synthesis problem. Rather than treating papers as monolithic files with rigid document-level tags, Lateral breaks PDFs down into granular paragraph-level "snippets" and organizes them inside a comparative matrix called the Table View. By pairing this matrix with semantic "Concepts" instead of standard keyword tags, researchers can train Lateral's AI on the fly to surface matching passages across their entire literature collection.

(Note: Lateral announced the sunset of its hosted SaaS platform on June 26, 2025, shifting its team's focus toward the decentralized open-science initiative coordination.network. However, its operational architecture remains a benchmark for how semantic AI and structured concept tagging accelerate systematic literature screening.)

1. Setting Up the Workspace and Importing PDF Collections

The foundation of an accelerated literature review in Lateral begins with constructing a project and importing your document corpus. Lateral provides two distinct ingestion workflows tailored to academic workflows.

Batch Uploading Local PDF Libraries

If you have already gathered a collection of full-text papers through databases such as Scopus, Web of Science, PubMed, or Google Scholar:

1. Open your project dashboard and click Add documents (either the primary button in the workspace or the header option).
2. Drag and drop your batch of PDF files into the upload window, or use the file selector to upload multiple files at once.
3. Click the blue Upload button to initiate ingestion.

During ingestion, Lateral automatically parses each PDF through its document processing pipeline:

• Snippet Decomposition: Lateral divides the text into logical, paragraph-sized units termed "snippets." This structural breakdown enables cross-document search and localized semantic suggestions.
• Bibliographic Metadata Extraction: The platform extracts core citation fields, including document title, author names, publication year, and journal source. Researchers can inspect and edit this metadata at any point by clicking the paper's title in the Table View.
• OCR Processing for Scanned Papers: If a paper consists of scanned image pages or non-selectable text, Lateral alerts the user and provides an integrated Use OCR option in the Document View to convert scanned text into searchable snippets using account OCR credits.
• Page Quotas: Ingestion consumes account page quota (e.g., standard Premium accounts provided 1,800 pages per month, or 21,600 pages under annual billing). Because page processing involves intensive machine learning parsing, deleting an imported paper does not restore used page credits.

Direct Ingestion via "Find Papers"

For expanding your library directly from within Lateral without manually downloading external PDFs:

1. Navigate to the Find Papers search tab.
2. Query the integrated Unpaywall database (which indexes over 30 million open-access scholarly articles) using keywords.
3. Refine queries using supported Boolean operators:
   • "Quoted text": Constrains the search to exact phrase matches within paper titles.
   • OR: Broadens queries to match either keyword (default search operates on AND).
   • - (hyphen/NOT): Excludes unwanted terms from article titles.
4. Click More info to inspect abstract details, then click the + icon on any paper to import and process it directly into your project.

2. Concepts vs. Tags: Why Semantic Tagging Matters

In conventional literature screening, researchers tag papers with labels like #methodology, #sample_size, or #adverse_events. This approach fails in two ways:

1. Lack of Precision: Tagging an entire 30-page paper does not help you locate the exact two sentences describing the dosage or participant criteria.
2. Keyword Rigidity: Standard keyword filters miss synonyms, alternate spellings, and related phrasing (e.g., searching for "side effects" misses "adverse drug reactions" or "toxicities").

The Mechanics of Lateral Concepts

In Lateral, a Concept functions as a dynamic semantic highlighter rather than a static tag:

• Passage-Level Grounding: Concepts attach directly to extracted text snippets or highlighted excerpts, not entire documents.
• Semantic Vector Representation: Lateral is powered by its proprietary Lateral Intelligence Platform (LIP) API—it does not rely on third-party GPT wrappers or share user document data for external model training. The platform understands synonyms, conceptual themes, and contextual phrasing.
• Paragraph-Length Concept Prompts: A Concept name does not have to be a single word. You can define a Concept using a detailed sentence or descriptive prompt (e.g., “Criteria used for excluding adult patients with secondary complications”). Lateral uses that full descriptive text to locate relevant passages.
• Cross-Project Reusability: Once configured, Concepts can be saved to your workspace's master Concept library and applied to subsequent review projects, retaining previous learnings.

3. Configuring Concept Tags Across Three Key Views

Researchers can configure and populate Concept tags using three primary entry points depending on their screening stage:

Method A: Configuring via Super Search (Broad Scoping)

When initially exploring an imported collection to see where a theme appears:

1. Open the Super Search bar in your project.
2. Enter your research question, key variable, or thematic phrase.
3. Scan through snippet results across all documents simultaneously.
4. On any highly relevant result, click the dropdown menu, type the name of your desired Concept, and select Create.
5. Lateral creates the Concept and immediately links that search result snippet to it.

Method B: In Document View (Deep Reading and Seeding)

When reading through a primary paper in Lateral’s built-in web reader:

1. Select a document from your library to open it in Document View.
2. Highlight the specific passage containing the empirical finding, variable, or methodology.
3. In the right-hand sidebar, open the Create or select concept dropdown.
4. Type a new concept name or select an existing one, then click Save.
5. Once saved, Lateral immediately evaluates that passage to begin recommending similar snippets across your remaining papers.

Method C: In Table View (Systematic Screening Matrix Setup)

When setting up a structured systematic extraction template before in-depth review:

1. Open the Table View (the primary workspace showing documents as rows).
2. Click the + button in the table header, or click the Concepts button.
3. Add the structured dimensions of your review as distinct Concept columns (e.g., Population & Eligibility, Intervention Design, Primary Measurement Tools, Risk of Bias / Limitations).
4. Your comparative extraction matrix is now configured and ready for rapid screening.

4. Accelerating Systematic Screening with Semantic Suggestions

Once you have defined your Concepts and saved just 1 or 2 anchor snippets to each, Lateral's active recommendation engine kicks in to expedite screening across the remaining literature.

Navigating the Table View Matrix

The Table View presents a two-dimensional grid:

• Rows: Every imported academic paper in the project.
• Columns: Your configured Concepts.
• Cells: The intersection between a specific paper and a specific research Concept.

Cell-by-Cell Semantic Suggestions

Instead of reading through paper 2, paper 3, and paper 4 from start to finish:

1. Click an empty cell in the Table View (e.g., the cell intersecting Paper 4 and Intervention Design).
2. Lateral automatically queries its semantic engine and displays Suggested Snippets in the sidebar. These snippets are paragraphs from Paper 4 that contextually mirror the snippets you previously saved under Intervention Design.
3. Click Save on relevant suggestions to accept them into the matrix, or dismiss irrelevant matches.
4. In Document View, the right sidebar similarly presents suggested snippets, allowing you to jump straight to relevant sections of a 40-page PDF without scrolling through unrelated introductory material.

Qualitative Flagging and Annotation

To maintain rigor during systematic screening:

• Color Flags: Lateral allows researchers to mark saved snippets with Green, Yellow, or Red flags. During screening, teams frequently use these flags to track inclusion eligibility, evidence confidence, or conflicting outcomes.
• Contextual Notes: Every saved snippet supports free-text notes. Researchers can add immediate paraphrasing, methodological critiques, or data points directly alongside the excerpted quote.

5. Exporting Findings and Completing the Screening Pipeline

A systematic review requires clear documentation and handoff into statistical software, writing platforms, or reference managers. Lateral provides several specialized export workflows from the Table View:

1. Microsoft Word (.docx):
   • Generates an organized dossier structured by Concept.
   • For every saved snippet, Word exports include the verbatim quote, inline citation data, and your attached qualitative notes.
   • Appends an automated, consolidated bibliography at the end of the document.
2. Microsoft Excel (.xlsx) & CSV:
   • Produces a comprehensive extraction spreadsheet where rows and columns map your screened findings.
   • Ideal for quantitative meta-analyses, PRISMA data tracking, or importing directly into Notion database templates.
3. BibTeX (.bib) for Reference Managers:
   • Exports all verified document metadata into standardized BibTeX format.
   • Allows seamless import into Zotero or Mendeley to ensure accurate citation keys when drafting the final manuscript.
4. Public Table Sharing:
   • Generates a read-only web link to your Table View matrix.
   • Enables research advisors, supervisors, or co-reviewers to inspect your screening findings and snippet matrices without exposing full underlying PDF files.

Practical Evaluation and Workflow Considerations

By breaking document silos into paragraph-level snippets, defining research questions as semantic Concepts, and leveraging cell-by-cell AI recommendations in the Table View, researchers can compress weeks of repetitive literature screening into structured, reproducible days.

References

• Lateral
• Lateral - Help Pages
• 2. Add Documents
• 3. Search across all your Documents
• 4. Save text to concepts
• 5. Get text suggestions
• Forget tags. Use Concepts to find text fast.
• How can I export my findings?
• What is Lateral?
• Terminology
• Does Lateral use GPT? No.
• How can I get more pages?
• Can I pay monthly? Yes.