Skip to content
Open app

Find the research nobody has done yet.

Gapmap collects new papers, models, datasets and code every day, links them to your research question, and shows the dataset × method combinations and blind spots that are still empty, with the evidence behind each one.

No account needed. Your studies stay in your browser.

Dataset × method coveragepapers per cell
SAM family × KiTS is emptyBoth are well studied, yet no paper in the corpus combines them. Gapmap lists the evidence on each side.
Pulls from OpenAlexEurope PMCarXivHugging FaceGitHubZenodoKaggle

Everything a scoping review needs before you start

Stop opening twelve tabs every morning. Gapmap keeps the literature, the models and the data for one question in one place.

Daily radar

New papers, models, datasets and repositories from the last day, week or month, ranked by how well they match your question.

One search, every source

Query journals, preprints, Hugging Face, GitHub, Zenodo and Kaggle together, with duplicates merged across sources.

Connection map

Datasets, methods and reported scores are pulled from abstracts and repository tags, then linked so you can see who used what.

Side-by-side comparison

Shortlist papers, datasets, models and code, then compare licenses, sizes, citations and reported scores on the same criteria.

Research ideas

Pick gaps and get a draft design: data, method, baselines, metrics and validation. Pro adds an AI-written proposal.

Reports and exports

A scoping report with numbered references, a BibTeX file and comparison tables, ready for your lab meeting.

Gap analysis is the core

Every gap is calculated from papers you can open, never guessed by a language model. A thinly studied area is not always an important one, so each card shows its evidence.

  • Untried combinationsDatasets and methods that are both active but never studied together
  • Blind spotsExternal validation, multi-center data, pediatric cohorts and more
  • ReproducibilityHighly cited papers that never shared code
  • Unused datasetsRepositories no paper has used yet, ready for external validation

From question to gaps in three steps

Describe your question

Write your research question and a few English keywords, or start from an example study.

Build a corpus

Gapmap gathers relevant literature, models and datasets since the year you choose, then keeps the radar updated.

Read the gaps

Open the heatmap and gap cards, add the ones you like to an idea, and export a report.

Calculations never use AI. When Pro writes a proposal, it may cite only papers from your corpus, and every citation number is checked automatically.

Simple pricing

Start free with one study. Upgrade when you need more studies, a longer radar, AI writing and exports.

Free

$0
For trying Gapmap on one question
  • 1 study
  • Radar for the last 7 days
  • Corpus of up to 25 items per source
  • Map, comparison and full gap analysis
  • Research idea drafts
  • JSON backup and restore
Start free

Pro Most useful

$9/ month
  • Everything in Free
  • Unlimited studies
  • Radar for the last 30 days
  • Larger corpus, set per source
  • AI search terms, summaries and proposals
  • Markdown, BibTeX, CSV and PDF exports
Get Pro

Prices in USD. Payments, tax and invoices are handled by Polar, our merchant of record. After checkout, Polar emails you a license key to paste into the app.

Questions

Do I need an account?

No. Your studies, shortlist and settings are stored in your browser. Export a JSON backup to move them to another device.

How do I unlock Pro after paying?

Polar emails you a license key. Open the app, go to Settings, paste the key under Plan, and Pro turns on right away.

Can I cancel any time?

Yes, from the Polar customer portal. Pro stays active until the end of the period you paid for.

Where does the data come from?

Public APIs from OpenAlex, Europe PMC, arXiv, Hugging Face, Zenodo, GitHub and Kaggle. Every item links back to its original source.

Can I trust the extracted numbers?

Datasets, methods and scores are extracted automatically from abstracts and repository metadata. Treat them as leads and confirm them in the full text before you design a study.

Your next study is probably sitting in an empty cell.