The Bibliomatica guide

1. What Bibliomatica is

Bibliomatica is a guided workspace for running a complete bibliometric study. A bibliometric study analyzes a body of published research to reveal what a field looks like: how it has grown, who its most influential authors and journals are, how its ideas cluster into themes, and where the gaps and frontiers sit.

Doing this well normally means stitching together several stubborn tools and losing weeks to manual work. Bibliomatica replaces that work with one continuous pathway of seven phases, from finding a topic to exporting a finished manuscript.

The problem it solves

A serious bibliometric review usually runs into three walls:

• The plumbing. Exports from different databases disagree on their columns, author names appear in a dozen spellings, and the corpus has to be merged, deduplicated, and cleaned by hand.
• The tooling gap. The standard analysis software is rigorous but silent, and it expects you to know R. The tools that write fluent text make the numbers up.
• The trust problem. When a general chatbot is asked to count citations or compute an h-index, it produces confident figures that are often invented. Those numbers can reach a reviewer with no way to catch them.

The core promise

Bibliomatica draws a hard line between two jobs.

• Verified software computes every number. Citation counts, h-indices, network metrics, and every statistic in your study are calculated by established research software, never by AI.
• AI writes only the words around those numbers. The AI drafts your methods, results, and discussion, but it is structurally unable to invent a figure. Numbers enter the text as placeholders bound to real values, and an automated check blocks delivery of anything it cannot trace to a source.

The result is a study you can defend line by line, where every claim is traceable and every step is reproducible.

Who it is for

Bibliomatica is built for anyone who needs a rigorous, reviewer-ready bibliometric analysis:

• Postgraduate students running a systematic literature review or a bibliometric thesis chapter
• Academics and research teams mapping a field for a review article
• Supervisors who want their students to produce defensible, reproducible work
• Any researcher who wants the speed of AI without the risk of fabricated evidence

What you can produce

By the end of the workflow you take home a full set of study documents, including:

• A registration-ready protocol and search strategy
• A merged, deduplicated, and cleaned corpus with a complete audit trail
• The full bibliometric analysis, with publication-standard figures and network maps
• A journal-ready Word manuscript with tables, figures, and a PRISMA flow diagram

2. How it works at a glance

Three movements

Everything you do in Bibliomatica fits into a story you can tell in one breath.

1. Plan. Find a research gap worth filling. Turn it into a registration-ready protocol and a search strategy.
2. Analyze. Merge your database exports into one corpus, clean it with AI assistance, and run the full analysis on established research software.
3. Publish. Draft the manuscript section by section, with every statistic pre-bound to a verified value, then export a Word document ready for submission.

The seven phases

Those three movements are delivered through seven disciplined phases. Each one produces real, named files that flow into the next.

The golden rule

One principle runs through the whole product: AI never computes a number. Deterministic research software does. Keep that in mind and the design of every phase makes sense.

3. What you need before you start

Bibliomatica runs entirely in your web browser. There is nothing to install. No R, no Java, no desktop software.

Here is the short checklist.

You need:

• A Bibliomatica account (a free sign-in, described below)
• A Pro subscription to run studies (see Plans and billing)
• A research question or topic area to explore
• Access to the databases you plan to search (for example a university subscription to Web of Science or Scopus)

You will want, at some point:

• A free API key for at least one data source. Some sources, such as OpenAlex, ask for a free key to use their data. Bibliomatica walks you through requesting one when a phase needs it.
• Access to an AI chat you already use, such as Claude, ChatGPT, or Gemini. Bibliomatica writes the prompts, and you run them in your own chat.

You do not need:

• Any statistics software installed on your machine
• Coding skills or knowledge of R
• A separate account for every tool. Bibliomatica brings the workflow together in one place.

A note on optional in-app assistance: a few steps can do work for you inside the app (for example, reviewing borderline duplicates or screening records). Those steps use your own AI provider key, stored securely in your vault. They are always optional, and you can complete a study without them by running the generated prompts in your own chat instead.

4. Getting started

Step 1. Create your account

Sign in with a secure sign-in link sent to your email, or with your Google account. There is no password to remember. Sessions are held securely on the server, so signing out cuts off access straight away.

Step 2. Set your access context

In Settings, choose the access context that matches how you use Bibliomatica. This is not a cosmetic preference. It decides which source is used to verify author identities during cleaning.

• Academic. Non-commercial research, the default for university and public-institution work. Enables author verification through ORCID's public data.
• Commercial. Industry or revenue-generating use. Author identity is verified through OpenAlex's open data instead, in line with ORCID's licensing terms.
• Institutional. Your institution holds ORCID or database memberships. You can add your member credentials to the vault for the richest data.

You can change this later if your situation changes, for example if you move from a university to industry.

Step 3. Add API keys when a phase asks for them

You do not have to set up everything up front. Bibliomatica tells you exactly when a key is needed and why. When a phase requires one, it points you to the right card in API Keys and, where relevant, gives you a one-click template to request a free key.

There are two kinds of keys, both explained in Core concepts:

• Data-source keys connect Bibliomatica to research databases and registries.
• AI-provider keys power the optional in-app assists.

Step 4. Create your first project

From the Projects page, select Create your first project. A project is one bibliometric study. Give it a topic area and a few starting details, and Bibliomatica opens the seven-phase workflow for you. Everything you produce, and every decision you make, stays scoped to that project.

5. Core concepts

A handful of ideas explain how the whole product fits together.

Projects

A project is one bibliometric study. Every output, every decision, and every audit record lives inside it. You can run as many projects as you like, and switch between them at any time.

The phase pipeline

The seven phases are shown as a pipeline. Each phase produces named files, and later phases need those files as their input. A phase stays locked until the phases before it have produced what it depends on.

This is deliberate. It means you cannot accidentally analyze a corpus you have not cleaned, or write up results you have not computed. The pipeline keeps your study in a valid order, and the trail of files is what makes the whole thing reproducible.

Your documents

As you work, Bibliomatica generates real files you keep: your protocol, your search record, your merged and cleaned corpus, audit logs, analysis outputs, figures, and your final manuscript. Nothing happens in a black box. You can see, download, and check every artifact. Understanding your outputs describes the main ones.

The credential vault

The vault is where your keys live. It holds two kinds:

• Data-provider keys for research sources such as OpenAlex, ORCID, ROR, Lens.org, and Dimensions.
• AI-provider keys for the optional in-app assists, from Anthropic, OpenAI, or Google.

Every key is encrypted at rest and only your signed-in session can unlock it. You manage all of this from the API Keys page.

Reporting standards, built in

Bibliomatica is designed for reviewer scrutiny. Four widely used reporting standards are woven into the workflow from the first step, not bolted on at the end:

• PRISMA 2020, the reporting framework for systematic reviews
• SPAR-4-SLR, a step-by-step protocol for systematic literature reviews
• Donthu et al. (2021), the standard guide for conducting bibliometric analysis
• TCCM, a framework for structuring a future-research agenda

Each is explained in the Glossary.

The verified-numbers guarantee

This is the founding idea, worth repeating. Every statistic in your study is computed by established research software or drawn from verified data files. During writing, numbers enter your manuscript as placeholders bound to those files, and an automatic audit blocks any figure it cannot trace back to a source. The AI writes the interpretation. It cannot invent the evidence.

6. Navigating the workspace

The workspace is simple to move around once you know the four areas.

The sidebar

On the left you will find:

• Projects. Your studies. Open one to see its phases and progress.
• Settings. Your account, your access context, and a summary of your vault.
• API Keys. Add, update, and remove your data and AI keys. Marked as your Vault.
• Help. A quick reference to the phases, shortcuts, and support.

The chevron on the sidebar edge collapses it to a slim icon rail when you want more room.

The top bar

Across the top:

• Project switcher. The project name is a switcher. Click it to jump between studies.
• Search. Find projects and phases from anywhere.
• Account menu. Your profile and sign-out.

The status line

Along the bottom, a status line always shows the active project, the current phase, your progress, and how many keys are in your vault. It follows you as you move through the phases so you never lose your place.

Shortcuts

• Command menu: press Command and K (or Control and K) to search projects and phases from anywhere.
• Escape: closes the command menu, any open menu, and dialogs.

7. The seven phases in detail

Each phase below follows the same shape: what it is for, what you do, what you need, and what you take away. The phases run in order, and each unlocks the next by producing the files it depends on.

A quick word on how the AI steps work throughout. In most phases, Bibliomatica writes a precise, study-aware prompt. You run that prompt in your own AI chat, copy the whole reply with the chat's Copy button, then paste it back into the app, which reads it into a structured form. This keeps your study out of anyone else's context and lets you use whichever AI you prefer. A few steps can also run inside the app using your own provider key.

Phase 1 — Foundation and Planning

What it is for. To choose a topic worth studying and turn it into a complete, registration-ready plan.

What you do. Find your research direction:

• I already have a topic. Start from the topic you gave the project. Bibliomatica hands you a validated Google Scholar Labs query for it; Scholar Labs reads recent primary studies and states, for each paper, what it tested, its main finding, and the gap it leaves open. You collect the gaps worth filling, each anchored to a citable source, and your assistant drafts the five sections of your research direction around them.
• Help me find a topic. Bibliomatica compiles an emerging-areas scan you run in your own AI chat, you pick one topic from its tables, and the same Scholar Labs gap search follows.
• Type the five sections myself. For prepared authors whose questions, objectives and references are already written. The form can still ask your assistant to draft whatever is missing, from exactly what you typed.

On the gap-search screen an optional novelty check protects the claim every bibliometric introduction depends on: that no map of your topic exists yet. One click generates three ready-made Google Scholar searches (AI widens the vocabulary; the searches still work without it), you run them in a new tab, and you record what you found. When the field is clear, a novelty gap card with the search date joins your collected gaps and carries the claim, with its provenance, all the way into your manuscript data.

Once you have a topic, Bibliomatica helps you generate your protocol and seal it section by section. Registration is a separate, optional step that follows: a bibliometric analysis has no registry requirement, unlike a systematic review, so you either prepare an OSF registration from the sealed protocol (and record its public link) or record that the study is not registered, and the search step opens either way. Then you test and refine your search strings against each live database, recording the result counts as you go. Sealing the strategy captures your search record (date and per-database counts), with a pre-filled Excel search record downloadable any time from the sealed step.

What you need. A topic area to start from.

What you get.

• Your complete protocol, search strategy, and search record as Word and Excel files you keep
• A recorded registration decision, and when you register, an OSF-ready registration built from your sealed protocol with its public link (INPLASY, Research Registry, and protocols.io to follow)
• A record of the gaps you selected, and of your novelty check when you ran one, carried forward to help write your introduction later

Phase 2 — Search

What it is for. To run your searches in the actual databases and bring the results back cleanly.

What you do. Follow step-by-step export walkthroughs for each database you committed to: Web of Science, Scopus, Lens.org, and Dimensions. Each walkthrough gives you the exact fields, formats, and settings to use so your export carries everything the later phases need. You then upload your raw export files, which are staged and validated before anything reaches your corpus.

What you need. Access to the databases you plan to search. The workflow adapts to the subscriptions you actually have, so you are never sent to a source you cannot use.

What you get. Your raw export files, checked and staged, ready to merge.

Phase 3 — Merge and Deduplicate

What it is for. To combine exports from different databases into one clean corpus and remove duplicate records.

What you do. Bibliomatica merges your files for you. Merging runs on Bibliometrix's own parsers, the same engine behind published bibliometric research, so your data is read the way the field reads it. A field-coverage matrix then shows you how complete your corpus is, and flags missing data before it can skew your analysis. It pays special attention to cited references, which several later analyses depend on.

Duplicates are found in seconds, right in your browser, by matching on DOI and on record metadata, with AI assistance offered on borderline pairs, such as a preprint versus its final version. Every decision is logged so you can review it, and the final deduplicated corpus and audit files are produced by the analysis service when you seal.

What you need. Your validated export files from Phase 2.

What you get.

• A single merged corpus
• A merge report describing coverage and any warnings
• A deduplicated corpus
• A deduplication audit that lists every removed record with its reason

Phase 4 — Filter

What it is for. To apply your inclusion and exclusion criteria and narrow the corpus to the records your study is actually about.

What you do. Set filters drawn directly from your protocol: document type, language, and year range. As you adjust them, a live counter shows how many records remain. For borderline cases, you can turn on an optional AI relevance screen. It is off by default and always ends with your review: the AI's flags come back in two groups, likely exclusions you spot-check and remove in one action, and uncertain records you judge one at a time. Nothing is removed until you confirm it. If you want to move on quickly, Apply all AI suggestions decides the remaining flagged records in one action, and the summary and the sealed record disclose that they were not reviewed one by one. Every removal is recorded with a reason.

What you need. Your deduplicated corpus from Phase 3.

What you get.

• A filtered corpus
• A filter log recording every decision. This audit trail is what lets your PRISMA flow diagram build itself later.

Phase 5 — Data Cleaning

What it is for. To fix the messiest part of any bibliometric study: the same author, institution, or keyword written many different ways.

What you do. Bibliomatica normalizes your author names, affiliations, and keywords against trusted reference sources: ORCID and OpenAlex for authors, ROR for institutions. Suggestions come in confidence tiers, so you can accept the obvious matches in bulk, review the ambiguous ones, and be sure nothing changes without your approval. Countries are extracted from address strings, and keywords are unified into a single controlled list.

Author names are resolved through OpenAlex author profiles for every study. Those profiles carry ORCID iDs from OpenAlex's open data, so ORCID's public API, which is licensed for non-commercial use only, is never called. Your access context (set in Settings) is recorded with your studies and does not change how names are matched.

The phase is five tabs, left to right: Audit (a read-only check that runs by itself), then Authors, Affiliations and Keywords (each runs a check and hands you a short list to confirm, in any order), then Seal. The Audit tab shows the plan with each step's state and a button for the next one. Nothing changes until you seal. On the Keywords tab the first thing you see is the spelling variants band: every case, punctuation, hyphenation, and plural variant the corpus contains, grouped under the form it uses most, with the rule that matched each one. Accept the whole band in one tap, untick the groups you want to keep apart, or decide group by group. Below it, the AI pass proposes merges that need judgment, such as synonyms and abbreviations, and those are reviewed the same way. The Seal tab counts anything you left undecided, and undecided keywords stay distinct in the cleaned corpus. On the Affiliations tab and on the Keywords tab, a long review list can be handed to your own language model: download the review packet, run the prompt in a new chat one part at a time, and paste the replies back. For keywords the model decides merge or keep per cluster and names the canonical form, and a merge is accepted only when that canonical is one of the cluster's own forms. Institutions the model names are checked in the Research Organization Registry before they are accepted, you see the counts before anything is applied, every decision stays reversible until you seal, and the cleaning report and Methods paragraph state what came from that channel.

What you need. Your filtered corpus from Phase 4, and a data-source key for author and affiliation lookups (Bibliomatica prompts you if one is missing).

What you get.

• A cleaned corpus
• Reusable normalization maps for authors and affiliations, which you can carry into future studies
• A keyword thesaurus compatible with network-mapping tools
• A cleaning report you can adapt straight into your methods section

Phase 6 — Bibliometric Analysis

What it is for. To compute the full analysis and build the figures and network maps that tell the story of the field.

What you do. Run the analysis and explore the results. Bibliomatica covers the full analytic canon, in two parts.

• Performance analysis: publication trends over time, source and author metrics, Lotka's law and Bradford's law, collaboration patterns, and citation impact.
• Science mapping: the conceptual and intellectual structure of the field, shown as interactive networks. This includes keyword co-occurrence, co-citation, bibliographic coupling, a thematic map, and thematic evolution over time.

Every number is computed by Bibliometrix on a managed cloud service. The AI computes nothing here. Network maps are rendered with VOSviewer, the tool reviewers recognize, and other figures use publication-standard color schemes so they drop straight into a manuscript. An inline filter lets you re-scope any figure on the fly, for example to look at a single date range, without changing your sealed corpus.

An optional AI step can help interpret and name the clusters your maps reveal. It works under strict guardrails: it will not label a cluster that is too small, every paper it cites must be a real record from your corpus, and it rates its confidence in each label.

What you need. Your cleaned corpus from Phase 5, and a data-source key where a step calls a live source.

What you get.

• The complete set of computed metrics as data files
• Interactive network maps you can explore and export
• Publication-ready figures and tables
• Optional written interpretations of your clusters and themes

Phase 7 — Writing and Reporting

What it is for. To turn everything you have produced into a journal-ready manuscript, with no invented numbers.

What you do. Bibliomatica assembles a single Writing Kit that bundles your verified numbers, your protocol, and your tables and figures for one continuous AI writing session. Guided prompts then walk you through drafting each part of the manuscript in order: a blueprint first, then Methods, Results, Discussion and Conclusion, Introduction and Abstract, and the front matter. Every statistic is pre-bound to a verified value, so the AI cannot invent a figure, and an automatic audit blocks anything it cannot trace to a source.

You polish the draft in a built-in editor that shows your manuscript taking shape section by section, then export it.

What you need. Your analysis outputs from Phase 6, and an AI chat capable of a long writing session.

What you get.

• A journal-ready Word manuscript, with your PRISMA flow diagram included
• A separate bundle of your tables and figures for submission
• A manuscript that meets the reporting standards reviewers expect: PRISMA 2020, SPAR-4-SLR, Donthu et al. (2021), and a TCCM-structured future-research agenda

8. Understanding your outputs

Bibliomatica produces real files at every phase. Here is what the main ones are and why they matter.

Corpus files. Your body of records as it changes through the study: merged, then deduplicated, then filtered, then cleaned. Each version is kept, so you can always see how the corpus evolved.

Audit trails. Plain records of every decision: which duplicates were removed and why, which records were filtered out, which name and affiliation mappings were applied. These are what make your study reproducible and what let you answer a reviewer's questions with evidence.

Reports. Readable summaries you can fold into your paper, such as the merge report and the cleaning report. The cleaning report, for example, is written to slot into your methods section.

Figures and tables. Publication-standard charts, network maps, and data tables, exported in formats ready for a manuscript.

Normalization maps and the keyword thesaurus. Reusable cleaning artifacts. The maps let you reuse your author and affiliation cleaning in a future study, and the thesaurus is compatible with network-mapping tools.

The manuscript. Your final Word document, assembled with verified numbers, a references section, and a PRISMA flow diagram.

Every number in these outputs comes from verified computation. None is produced by AI.

9. How Bibliomatica uses AI

Bibliomatica is careful and deliberate about AI, because that is the whole point of the product.

AI writes. Software computes. The AI is allowed to do one thing: write the narrative around values that verified software has already computed. It never produces a statistic of its own.

You use the AI you already have. For most steps, Bibliomatica writes a precise, study-aware prompt. You run it in your own chat with Claude, ChatGPT, or Gemini, copy the whole reply with the chat's Copy button, and paste it back. Your study never shares a context window with anyone else's work.

Optional in-app assists use your own key. A few steps can run inside the app, such as reviewing borderline duplicates, screening records, and disambiguating names. These use your own AI-provider key from Anthropic, OpenAI, or Google, stored encrypted in your vault. They are always optional.

The numbers are protected. During writing, every figure enters the text as a placeholder bound to a real value in a real file. An automatic audit runs before delivery and blocks any number it cannot trace to a source. This is what makes a Bibliomatica manuscript defensible where general AI tools are not.

10. Data, privacy, and security

Your work and your credentials are handled with care.

Your keys are encrypted. Every key in your vault is encrypted at rest with strong encryption, and only your signed-in session can unlock it.

Sessions are server-side. Your session is held on the server, not just in your browser, so signing out revokes access immediately.

Your corpus is private. Your uploaded files and your corpus are stored privately to your project.

Your work is never used to train AI models. Full stop.

Access and licensing. Author names are resolved through OpenAlex's open (CC0) author profiles for every study. Those profiles embed ORCID iDs, so ORCID's public API, which is licensed for non-commercial use only, is never called. The access context you set in Settings is recorded with your studies and shown on the connections that depend on it, such as Lens.org.

11. Plans and billing

Bibliomatica has one simple plan.

Pro. Everything you need to complete a bibliometric study from foundation to a submission-ready manuscript. It includes:

• Unlimited projects
• The full seven-phase guided pipeline
• Uncapped fair-use analysis jobs
• Metrics computed by deterministic software, with AI writing only around verified values
• Submission-ready manuscript export, with tables, figures, methods, and the PRISMA flow
• The full Phase 7 Writing Kit
• The option to bring your own OpenAI, Anthropic, or Google key
• Email support from the Bibliomatica team

Price. RM49 per month for Malaysian users, or USD 15 per month for international users. Billed monthly, at launch. You can cancel anytime.

Payment. Malaysian users pay through CHIP (FPX and QR). International users pay by card through Stripe.

Refunds. All sales are final, so you are encouraged to preview the complete workflow in the tutorial before subscribing. Review the refund policy before you purchase. You manage your subscription, renewal, and payments from the Billing area in Settings.

12. Tips and troubleshooting

A few things smooth the path.

Add a data-source key early. Cleaning in Phase 5 needs one. Request your free key at the start so you are not waiting on it later. Bibliomatica gives you a one-click request template.

Export the right fields in Phase 2. Each database card lists the export settings to use, and the file type and the field groups you tick decide how complete your corpus is. The most common cause of a weak analysis is an export that left out cited references. Bibliomatica pauses before staging a Scopus or Web of Science file that has none, and Phase 3 flags it again, but it is easier to export correctly the first time.

Watch the field-coverage matrix in Phase 3. If a critical field like cited references is missing, several Phase 6 analyses will be limited. The matrix tells you before you rely on the data.

Clean in confidence tiers. In Phase 5, accept the clearly correct matches in bulk and spend your attention on the ambiguous ones. Nothing changes without your approval, so you stay in control without reviewing every single record by hand.

You can step away while a server step runs. Merging, the corpus audit, author and affiliation resolving, the keyword inventory, every seal, and every Phase 6 analysis run on Bibliomatica's servers, not in your browser. Switch tabs, close the page, or let your laptop sleep: the step finishes on its own and its result is waiting when you come back. Each of these running panels has an "Email me when it's done" button, so you can leave and get one email with a link back when the step finishes or hits a problem. AI screening is different: it runs from your browser in batches, pauses when you leave, and resumes where it stopped when you return. The author, affiliation and keyword AI passes also run from your browser, but their panels carry the same "Email me when it's done" button: tick it and the pass finishes on Bibliomatica's servers if you close the tab, then one email tells you it is done.

Use a long-context model for writing. Phase 7 is one continuous session. A writing model that holds a lot of context will keep your sections consistent from Methods through to the Abstract.

The software is fast. The method takes judgment. Merging, deduplication, and analysis run in minutes. What takes time is the review the method demands: screening records, confirming matches, and approving each phase. Bibliomatica removes the weeks of spreadsheet work between those judgment calls, not the judgment itself.

13. Frequently asked questions

Do I need to install R, Java, or anything else? No. Everything runs in your browser. The statistical engine runs on a managed cloud service, and the interactive network maps are built in. Desktop tools remain optional extras for advanced users, never a requirement.

What does it cost? Pro is RM49 per month for Malaysian users and USD 15 per month for international users, monthly only at launch. All sales are final, so preview the full pipeline in the tutorial before subscribing.

Do I need my own AI subscription? Bibliomatica works with the AI you already use. Each phase writes precise, study-aware prompts that you run in your own chat with Claude, ChatGPT, or Gemini, so your study never shares a context window with anyone else's. Optional in-app assists use your own key from Anthropic, OpenAI, or Google, stored encrypted in your vault.

Which databases are supported? Web of Science, Scopus, Lens.org, and Dimensions for record exports. OpenAlex, ORCID, and ROR provide trusted reference data for topic discovery, author verification, and affiliation cleaning. Bibliomatica asks about your database access up front and adapts the workflow to what you have.

Can I trust the numbers in the manuscript? Yes, and that is the founding principle. Every statistic is computed by verified software or drawn from verified data files, never by AI. During writing, numbers enter the text as placeholders bound to those files, and an automated audit blocks delivery of any figure it cannot trace to a source.

Is my data secure? Your keys are encrypted at rest, and only your signed-in session can unlock them. Sessions are stored on the server, so signing out revokes access at once. Your corpus is stored privately, and your work is never used to train AI models.

How long does a study take? The software is rarely the bottleneck. Jobs run in minutes. What takes time is the review the method requires: screening records, confirming ambiguous matches, and approving each phase. Bibliomatica removes the weeks of manual work, not the judgment.

What is the difference between academic and commercial use? Nothing in the cleaning itself. Every study resolves author names through OpenAlex's open (CC0) author profiles, which embed ORCID identifiers, so ORCID's non-commercial public API is never called. Your access context is recorded with your studies and shown on the connections that depend on it, such as Lens.org.

14. Glossary

Bibliometric analysis. The quantitative study of published research. It measures things like publication and citation patterns, and maps how the ideas and authors in a field relate to one another.

Corpus. The collection of research records your study is built on, after you have merged, cleaned, and filtered it.

Deduplication. Finding and removing records that appear more than once, for example the same paper exported from two different databases.

PRISMA 2020. A widely used reporting framework for systematic reviews. It sets out what to report and includes the PRISMA flow diagram, a chart showing how many records you found, screened, and included.

PRISMA flow diagram. The standard chart that traces records from the initial search through screening to the final included set. Bibliomatica builds yours from your filter audit trail.

SPAR-4-SLR. A step-by-step protocol for conducting and reporting a systematic literature review, used to structure the method.

Donthu et al. (2021). A widely cited guide, "How to conduct a bibliometric analysis," that sets out the accepted techniques and good practice for this kind of study.

TCCM. Theory, Context, Characteristics, and Methodology. A framework for organizing a future-research agenda in a review, so your discussion suggests next steps in a structured way.

Performance analysis. The part of a bibliometric study that measures productivity and impact, such as which authors, journals, institutions, and countries publish and are cited the most.

Science mapping. The part of a bibliometric study that reveals the structure of a field by mapping relationships between its documents, keywords, and references.

Keyword co-occurrence (co-word analysis). A map built from keywords that appear together, showing which topics cluster and how the field's themes are organized.

Co-citation analysis. A map built from papers that are cited together, revealing the intellectual foundations a field draws on.

Bibliographic coupling. A map built from papers that cite the same earlier work, useful for spotting current research fronts.

Thematic map. A chart that plots a field's themes by how developed and how central they are, so you can see which are established, which are emerging, and which are niche.

Thematic evolution. A view of how a field's themes appear, merge, and fade across time periods.

Lotka's law. A classic pattern in bibliometrics describing how a small number of authors produce most of the publications in a field.

Bradford's law. A classic pattern describing how a small core of journals accounts for a large share of the articles on a subject.

h-index. A measure combining productivity and impact. An author has an h-index of h if they have h papers each cited at least h times.

Bibliometrix. Established research software that computes bibliometric analyses. In Bibliomatica it runs on a managed cloud service, so you never install it.

VOSviewer. A widely recognized tool for building and viewing bibliometric network maps. Bibliomatica embeds it so your network figures look the way reviewers expect.

OpenAlex. A large open catalogue of scholarly works, authors, and institutions. Bibliomatica uses it for verified author and affiliation data. Some of its features ask for a free API key.

ORCID. A registry of unique researcher identifiers. Bibliomatica uses it to verify author identities during cleaning.

ROR (Research Organization Registry). An open registry of research institutions. Bibliomatica uses it to normalize the many different ways an institution's name can be written.

Web of Science, Scopus, Lens.org, Dimensions. Major research databases you search and export from. Bibliomatica gives exact export instructions for each.

API key. A private access code that connects Bibliomatica to a data source or an AI provider on your behalf. Keys live encrypted in your vault.

Access context. Your declaration of how you use Bibliomatica (Academic, Commercial, or Institutional). It decides how author identities are verified, in line with each provider's licensing.

Credential vault. The secure store for your keys, both data-source keys and AI-provider keys.

Protocol. The plan for your study: its question, aims, criteria, databases, and search strategy. Bibliomatica helps you produce a registration-ready one.

Registration. Publishing your protocol on a platform such as OSF before you search. Expected of systematic reviews; optional for a bibliometric analysis, which no registry or reporting guideline requires to be registered. Bibliomatica records your decision either way.

Writing Kit. The single bundle Bibliomatica assembles for Phase 7, containing your verified numbers, protocol, tables, and figures, so one continuous AI writing session has everything it needs.

15. Getting help

If something is unclear or a step is blocking your study, contact the Bibliomatica team. Study-blocking questions are answered first. Chat is fastest. Use email for anything long.

WhatsApp: +60 17-470 4553

Messenger: m.me/workcyte

Email: support@bibliomatica.app

Inside the app, the Support button in the top bar opens the same three channels. You can also open Help for a quick reference to the phases, keyboard shortcuts, and answers to common questions.

Bibliomatica gives every researcher a bibliometric study they can defend line by line. The language work goes to AI. The arithmetic goes to verified software. From a first question to a submission-ready manuscript, in seven reproducible phases.