FREE TOOL · NO SIGN-UP
Free dataset finder
Search research datasets in Zenodo, Dryad, Harvard Dataverse, Hugging Face and DataCite in one go.
Free: 5 runs a day per IP address, shared across the free tools. No account. Returns up to 6 datasets per search; queries are cut at 200 characters.
Input
Results
Press Run to see live results here. See a saved sample result below
How the dataset finder works
- 1
Describe the data
Type a topic or a few keywords, such as “urban tree canopy”. Short, specific phrases work best.
- 2
Five sources are searched
The query is sent to Zenodo, Dryad, Harvard Dataverse, Hugging Face and DataCite at the same time. A dataset that turns up twice, by DOI or by title and year, is merged into one result.
- 3
Read the results
You get up to 6 datasets, taken in turn from each source so that one repository does not fill the list. Each shows its repository, year, first creators and, when the repository states one, the licence.
Sample result from the dataset finder
Saved from a real run of this tool on 13 October 2026, using the example input that is pre-filled above. The lookup took 1.5 seconds in that run. Live results can differ as the databases change.
Example input
urban tree canopy
Three things in this run are worth noticing. The figshare record arrived through DataCite. The Harvard Dataverse record states no licence, so none is shown. And Hugging Face answered but had no dataset for this query.
Saved result
6 datasets for "urban tree canopy" from zenodo, dryad, dataverse, huggingface, datacite.
Dryad · 2021 · Danica Doroski
Harvard Dataverse · 2026 · Guignard, Quentin
Dryad · 2022 · Jacqueline Buenrostro; Ruth Hufbauer
Which repositories are searched
Five sources, all through their public APIs:
- Zenodo: a general repository for research outputs in any field.
- Dryad: research data, usually deposited alongside a journal article.
- Harvard Dataverse: a repository open to researchers in all disciplines.
- Hugging Face: datasets for machine learning.
- DataCite: the DOI registry used by many other repositories, which is how records from places such as figshare appear in the results.
A licence is not the same as access
The licence on a result is the dataset’s own licence when the repository states one. It tells you what you may do with the data once you have it. It does not tell you that the files are open to download.
The first record in the sample shows the gap: Zenodo lists it under CC-BY-4.0, and the same record marks its files as restricted access. Open the record and check the access conditions before you plan a project around a dataset. A result with no licence label means the repository did not state one.
How results are merged and ordered
The same dataset is often registered in more than one place, for example in Zenodo and in DataCite. Results are matched by DOI first, then by title and year, and the repository’s own record is kept because it usually has more detail.
The list then takes one result from each source in turn. With 6 results on this page, that gives a spread across repositories, not the 6 best matches from any single one. To go deeper into one repository, search it directly.
FAQ
Related tools
- NIH RePORTER search by topic
See which NIH-funded projects work on the same topic.
- Free citation checker
Check that every reference in your list exists and is cited correctly.
- Dataset search for AI agents
The same search with more results, sizes, formats and access status.
Data sources behind this tool: Zenodo data source, Dryad data source, Harvard Dataverse data source, Hugging Face Datasets data source, DataCite data source. All ScholarMuse web tools
The same search, inside Claude or ChatGPT
Connect ScholarMuse once and your agent can search the same five sources, read the descriptions and shortlist datasets that fit your question. Calls from an agent need an account and use credits.