Zenodo Research Search Scraper

Search Zenodo’s public research repository by keywords and filters, then save normalized publication, dataset, software, image, video, poster, presentation, lesson, or other research records to an Apify dataset.

The Actor uses Zenodo’s public records API. It does not require a Zenodo account, download deposited files, access private records, solve CAPTCHAs, or bypass access controls.

What it extracts

Each record can include:

• Zenodo record and concept identifiers, DOI, canonical record URL, API URL, preview URL, and version URLs
• title, plain-text description, publication/creation/update dates, resource type, access right, embargo information, language, and version
• normalized creators and contributors, affiliations, ORCID/GND identifiers, keywords, subjects, references, and notes
• license, related identifiers, journal/conference/imprint/thesis metadata, grants, and communities
• normalized file metadata (name, byte size, checksum, public API/download URLs) without downloading the files
• views/download statistics, result page, total search result count, query, sort order, and collection timestamp

Raw API responses are not stored. Empty values are recursively removed, records are deduplicated by Zenodo record ID, HTML descriptions are converted to plain text, and only HTTP(S) URLs are emitted.

Input

The Actor opens the public Zenodo search API using the values you would enter or select in Zenodo’s search interface.

Field	Purpose	Default
query	Keywords or a Zenodo advanced-search expression	empty
resourceType	Publication, dataset, software, poster, presentation, image, video, lesson, or other	publication
accessRight	Open, closed, embargoed, or restricted	all
community	Exact Zenodo community identifier	all
startDate, endDate	Inclusive publication date range (YYYY-MM-DD)	unbounded
sortBy	mostrecent or bestmatch	mostrecent
maxItems	Maximum unique records	20
maxPages	Pagination safety limit	10
pageSize	Stable records-per-page request size; writes are separately capped by maxItems	25
maxRequestRetries	Temporary-failure retries	2
requestTimeoutSecs	Timeout per API request	30
requestDelayMillis	Polite delay between pages	500
useApifyProxy	Use Apify Proxy	false
proxyGroups, proxyCountryCode, proxyUrls	Optional proxy configuration	unset

The deprecated limit field remains accepted for backward compatibility, but cannot be combined with maxItems.

Example

{
  "query": "machine learning",
  "resourceType": "dataset",
  "accessRight": "open",
  "startDate": "2024-01-01",
  "sortBy": "bestmatch",
  "maxItems": 50,
  "maxPages": 5
}

No-results searches finish successfully with an empty dataset. Invalid inputs fail before any request is sent.

Run locally

Install dependencies and run through the Apify CLI:

npm ci
apify run --purge --input '{"query":"open science","maxItems":10}'
npm run validate:local

Run unit tests and schema validation with:

npm test
apify validate-schema

Run on Apify Cloud

Deploy with apify push, then call the mapped Actor with a bounded JSON input. Results are written to the run’s default dataset. Public Store visibility, pricing, and monetization are separate settings and are not changed by deploying source code.

Reliability and limitations

The Actor validates HTTP status and JSON content type before parsing, rejects malformed payloads, retries only temporary network/rate-limit/server errors with bounded exponential backoff, and supports direct, Apify Proxy, or custom proxy connections. Requests are sequential and rate-limited.

Results depend on Zenodo indexing, public API availability, search syntax, and rate limits. Optional metadata legitimately varies by deposit. File links and metadata are returned, but file contents are not downloaded.