AI Research Gap Finder & Literature Search

A research gap finder surfaces open, unanswered questions in a field so you can pick a defensible, novel direction. Ours works literature-first: a free search returns the field overview (queried live against 250M+ OpenAlex works), a comparison table that lists the papers closest to your topic first, and — below the gaps — every paper the search found in one list you can filter, sort and download as RIS or BibTeX; unlocking that topic (50 credits) reads the papers across 120,000+ open gaps mined from 100,000+ papers in our local library and synthesises the gaps from what the filled rows actually say.

Yes, and searching is free here: browsing our 1,100+ pre-analysed research gaps by field needs no account at all, and with a free account you can run the live search itself — the field overview, the comparison table with the papers closest to your topic listed first, and, below the gaps, one list of every paper the search found that you can filter, sort and download as RIS or BibTeX — for no credits. The one paid step is unlocking the topic (50 credits), which is what produces the research gaps, the cross-paper synthesis, the proposal and the table's downloads. A new account starts with 25 welcome credits and a referral adds 25 to each side, so 25 + 25 covers one whole topic.

Enter a topic or paste your abstract — the search is free. You get a real-time OpenAlex field snapshot (total works, top venues, year range and peak year), a comparison table that lists the papers closest to your topic first (title, year and venue filled, its field columns named but empty), and — below the gaps — every paper the search found in one list, each with its year, venue, journal quartile and citation count where known. Narrow that list by year, journal quartile, indexing, which list found a paper (closest match, most cited or most recent) and venue, sort it, and download what is left as RIS or BibTeX, all free. Unlocking the topic costs 50 credits and searches our whole collection: objectives, methods, variables, datasets, results, limitations and stated gaps side by side, with the table's downloads included (its grid as CSV, XLSX or PDF, its own papers as RIS or BibTeX) — and the research gaps themselves, synthesised from those filled cells so each one is traceable to a row you can read. Full-sentence queries retrieve more specific literature than keywords.

Searching costs nothing: creating an account is free and comes with 25 credits to start, but the field overview, the empty comparison table and the list of every paper the search found — with its filters, its sorting and its RIS and BibTeX download — spend none of them. One topic costs 50 credits, once, and that single charge covers everything the gap tool does for that topic. Every paper in the table is read for 21 named fields — objectives, problem statement, methods, dependent and independent variables, dataset, population sample, results, findings, limitations, each paper's own stated research gap, its future-research note and the rest. Unlocking searches our whole collection for your topic, so it adds papers beyond the ones the free search listed and keeps the closest of those pinned at the top. You also get the cross-paper synthesis, the research gaps written from those same rows, the Research Proposal drafted on them, and one Pro Report run over those gaps. The table's downloads come with it — CSV, XLSX and PDF rendered from the same grid you are looking at, and RIS and BibTeX of the table's own papers for your reference manager — and re-downloading costs nothing, ever. A column no delivered paper fills is left out rather than shown empty; rows come from a paper's full text where we hold it and from the abstract where we do not, and every row is labelled which. Credits are bought in one-time packs and never expire; a referral gives 25 credits to each side, so 25 welcome + 25 referral is exactly one topic.

Because the rest is made of the table. The gaps are synthesised from the rows the fill has just written, the proposal is drafted on those same rows and gaps, and the Pro Report goes deeper over them — there is no second search and no second pass over a different corpus to charge for, and that is exactly why each gap can point at a row you can read instead of asking you to trust it. Charging again for something computed from what you already bought would be charging twice. So the topic comes as one package: the filled table, the synthesis, the gaps, the proposal and a Pro Report run, for 50 credits, with nothing inside it sold separately. Re-generating the proposal for that topic is free as well.

You are not charged. A comparison needs a minimum number of papers we hold comparable extractions for; below that the fill is refused before any credits move — a refusal, not a refund — and the page tells you the count it found. The same applies if we cannot verify the retrieved papers are actually on your topic. If a table is built but fails to save, the credits are returned automatically, and if that return itself fails we say so rather than tell you it landed. Two limits we will not dress up: we can only compare papers we hold a structured extraction for, so a very new or very narrow topic can fall under the floor — and the gap pass can come back empty on rows too sparse to compare, in which case you keep the full table and the table says so.

It combines real bibliometric evidence (OpenAlex coverage, citation and recency signals) with gaps mined from the abstracts, limitations, future-work and other gap-stating passages of 100,000+ papers, and surfaces each with the source behind it — grounded in real literature, not invented.

Yes. Use the candidate gaps as starting points, confirm each by reading the 'future work' of the surfaced papers, then run a Pre-Check or full AI Review once you have a draft.

Yes — enter the topic or paste an abstract and the finder returns the literature on exactly that subject, free: a comparison table built for it with the closest papers listed first, and below the gaps every paper it found, in one list you can filter, sort and download as RIS or BibTeX. You can deep-link a topic directly as /research-gaps?topic=<your topic>. Unlock that topic for 50 credits and you get the evidence-backed open questions for the subject, each grounded in a cell of the table, plus the proposal and a Pro Report run on the same topic. A free account starts with 25 welcome credits.

Yes, free. Below the gaps, every paper the search found — the closest matches, the most cited and the most recent, merged into one list with duplicates removed and each marked with the lists that found it — can be searched by title, venue or author, narrowed by publication year, journal quartile, where the journal is indexed (Web of Science, Scopus, PubMed, DOAJ or TR Dizin), which list found it and venue, and sorted by best match, citations, journal quartile, date or title. The filters work over the papers that search returned, not over all of OpenAlex; to change the pool, change the query. A filter none of those papers has data for is left out rather than shown empty. The .RIS and .BibTeX buttons download every paper that passes your filters, in the order you sorted them, not only the ones on screen. Citation counts and quartiles are shown only where they are known: a paper from our collection with no recorded count says so instead of showing zero. That download is separate from the comparison table's, which comes with unlocking the topic and exports the table itself — its grid as CSV, XLSX or PDF, and its own papers as RIS or BibTeX.

Yes. Every search runs a real-time query against 250M+ works in the OpenAlex API for field coverage, top venues, and citation/recency signals, combined with 120,000+ open gaps mined from 100,000+ papers in our own library.