University Library Bern · Digital Scholarship

Deutsch

Data for TDM and AI: Access and Conditions

Interfaces, formats and contract terms at a glance.

As of 3 November 2026 · 20 publishers and 5 databases

Important: The information on contract terms is for rough orientation only and does not constitute legal advice. Contract terms take precedence over information on provider websites. Please clarify any TDM or AI use of licensed content in advance with Digital Scholarship (ds.ub@unibe.ch).
permitted conditional not permitted without an additional agreement not specified not covered by the contract Crossref TDM links for ≥ 80 % of articles

Overview

Access:
Data:
ProviderTDMAIAccessFormatsQuota
Publishers
ACM
ACM Digital Library		permittednot specifiedOpen Access; API per contract, no public documentation–no volume limit
ACSnot permittednot permittedno API; separate contract onlyJATS XML (journals), BITS XML (books), by agreement–
AIP Publishingpermittednot specifiedAPI per contract, no public documentation–as specified by AIP
APS
Physical Review		conditionalconditionalHarvest API; bulk data sets (metadata, citations) on applicationJSON, XML, PDF, BagIt–
CUP
Cambridge Core		conditionalnot specifiedno API; download via Cambridge CorePDF500 PDFs/hour
Elsevier
ScienceDirect		conditionalpermitted
closed tools		ScienceDirect APIs; API keyXML, JSON, Atom, PDF, TXT
(tool: CSV, JSON, XML, TXT)		50,000 full texts/week
Frontierspermitted
CC BY		permitted
CC BY		no API; download from the article pagesPDF, XML; EPUB (e-books)–
IEEE
IEEE Xplore		conditionalconditionalAPI key; full text with authorization keyJSON, XML10 requests/s, 200/day
IOP Publishingconditionalconditionalno API; delivery on request via sFTP or another agreed routeXML, PDF–
Kargerpermittedpermittedno API; TDM after request; FTP data service per website (for industry customers)XML, PDF–
LWW
Wolters Kluwer Health		conditionalnot specifiedno API––
MDPIpermittedpermittedOAI-PMH open; FTP on requestDublin Core, AGRIS; PDF, XML–
OUP
Oxford Academic		conditionalnot specifiedno API; further access after whitelisting––
PLOSpermitted
CC BY		permitted
CC BY		Search API without key; complete download allofPLOS; also via PubMed CentralXML, JSON; JATS XML7200 requests/day, 10/min
RSCconditionalnot specifiedno API; delivery route by agreement––
Sagenot specifiednot specifiedno API; via Crossref linksas per Crossref link1 request every 2–6 s
Springer NatureconditionalconditionalAPI key, TDM key on requestJATS XML, JSON, XML
(tool: CSV, JSON, JATS XML, TXT)		500 hits/day
Taylor & Francisconditionalnot specifiedno API; register project in advance––
UC Pressnot specifiednot specifiedno API; only individual titles licensed––
WileyconditionalconditionalTDM token (click-through)PDF3 requests/s
Databases and metadata services
Crossrefpermitted
metadata		not specifiedREST API without key; mailto recommended; Public Data Files (dump)JSON5–10 requests/s
Dimensionsconditionalnot specifiedAnalytics API; API key, used to obtain a token (valid for about 2 hours)JSON30 requests/min
OpenAlexpermitted
CC0		permitted
CC0		REST API; free key recommended; data dump (quarterly)JSON10–30 requests/s
PubMed Centralpermittedconditional
depending on licence		Cloud (AWS), OAI-PMH, E-Utilities, BioC APIJATS XML, TXT, PDF, BioC (XML, JSON); Dublin Core–
Scopusnot specifiedpermittedScopus APIs; API keyJSON, Atom, XML20,000 searches/week

Details by publisher

Association for Computing Machinery (ACM)TDM permittedAI not specifiedCrossref partly: 39 % / 93 %

Key links

API key

  • no public API documentation; per contract, API provided by ACM or an agreed third-party provider

Content, formats and quotas

  • ACM Digital Library fully Open Access since 2026, with basic TDM functionality
  • Per contract at no additional cost and without volume limit

Contract terms (excerpt)

  • TDM for academic research and education
  • TDM results may be shared as long as no substitute product for third parties is created
  • No robots or agents outside the designated channels

Crossref: text-mining links for 39 % of recent and 93 % of older articles; licence information for 58 % and 94 % respectively.

American Chemical Society (ACS)TDM not permittedAI not permitted

Key links

API key

  • no public API

Content, formats and quotas

  • TDM for a fee by agreement; sample files in JATS XML (journals) and BITS XML (books)
  • AI Summarization License as an add-on for AI summaries; no training of language models

Contract terms (excerpt)

  • Data mining, machine indexing and aggregation only with a separate agreement
  • No bulk downloads, no central storage
  • Individual articles for personal research and teaching only

Crossref: no text-mining links; licence information for 94 % of recent and 19 % of older articles.

AIP PublishingTDM permittedAI not specified

Key links

  • no public API documentation

API key

  • TDM only via an API service approved by AIP

Content, formats and quotas

  • No public information on formats
  • Load limits and download rates as specified by AIP

Contract terms (excerpt)

  • Scholarly purposes only
  • Storage on the secure network or another secure location; results not to be made publicly available on the internet
  • No systematic downloading outside the API

Crossref: text-mining links for 15 % of recent and 4 % of older articles; licence information for 34 % and 5 % respectively.

American Physical Society (APS)TDM conditionalAI conditional

Key links

API key

  • Access via CHORUS token, API key or the university's IP range
  • Open Access articles without authentication

Content, formats and quotas

  • Harvest API: article lists and individual articles by DOI; APS Article JSON, XML, PDF or BagIt package (full-text XML, metadata, figures, PDF)
  • APS Data Sets: metadata and citation pairs as a complete data set, on application

Contract terms (excerpt)

  • TDM only on written request; APS may not unreasonably withhold permission
  • Research, teaching and private study only, non-commercial; TDM also covers AI
  • AI only locally, self-hosted or in a closed environment; training and fine-tuning only with written consent
  • No crawling, no systematic permanent copies

Crossref: no text-mining links; licence information for 100 % of recent and 99 % of older articles.

Cambridge University Press (CUP)TDM conditionalAI not specified

Key links

API key

  • no API; download via the Cambridge Core website

Content, formats and quotas

  • Download, extraction, storage and indexing for TDM permitted
  • Max. 500 PDFs per hour

Contract terms (excerpt)

  • Delete copies once the analysis is complete
  • Results only on personal devices or on the university's secure network
  • No crawling except for federated search and discovery

Crossref: text-mining links for 3 % of recent and 5 % of older articles; licence information for 100 % and 97 % respectively.

Elsevier (ScienceDirect)TDM conditionalAI permittedCrossref 100 % / 98 %
Download tool: Elsevier Full-Text Downloader, runs in the browser, retrieval by DOI or PII.
  • Metadata with abstracts: CSV, JSON
  • Full texts: XML (original), TXT
  • Licensed full texts only on the campus network or via VPN

Key links

API key

  • Account on the Developer Portal and API key

Content, formats and quotas

  • ScienceDirect Search API: metadata; 20,000 requests/week, 2 per second, max. 200 hits per query
  • Article Metadata API: metadata with abstract (COMPLETE view); JSON, Atom, XML; max. 6000 results
  • Article Retrieval API: full texts as XML, JSON, PDF, TXT; 50,000/week, unlimited with a text-mining key at 10 requests/s

Contract terms (excerpt)

  • University members only, academic research only, non-commercial
  • AI use in a closed AI tool permitted, including training
  • Delete the data set after the project ends; no crawling

Crossref: text-mining links for 100 % of recent and 98 % of older articles; licence information for 100 % and 98 % respectively. The Article Retrieval API also accepts a Crossref token.

FrontiersTDM permittedAI permitted

Key links

API key

  • no API; freely accessible

Content, formats and quotas

  • Download from the article pages as PDF and XML (e-books also as EPUB)

Terms of use

  • No contract; content under CC BY: use with attribution

Crossref: hardly any text-mining links (0 % / 3 %); licence information for 100 % of recent and 75 % of older articles.

IEEE (IEEE Xplore)TDM conditionalAI conditional

Key links

API key

  • Registration and API key
  • Full texts: additional authorization key from IEEE

Content, formats and quotas

  • Metadata Search API: metadata with abstracts; JSON or XML; max. 200 hits per query, up to 25 DOIs per request
  • Open Access API / Full-Text Access API: full texts by article number; JSON or XML
  • Max. 10 requests per second and 200 per day

Contract terms (excerpt)

  • Notify every TDM activity in advance; the access route is agreed jointly
  • AI only non-commercially in a local or closed environment, without training
  • More than 100 documents in an AI tool only after an application via the API

Crossref: no text-mining links; licence information for 99 % of recent and 96 % of older articles.

IOP PublishingTDM conditionalAI conditionalCrossref partly: 100 % / 60 %

Key links

API key

  • no public API; delivery on request

Content, formats and quotas

  • Request with a DOI list or publication period; delivery via sFTP
  • XML metadata (including abstracts and references); full text as PDF for almost all articles, as XML for most recent articles

Contract terms (excerpt)

  • Internal and non-commercial use only; no crawling
  • AI for internal purposes only, otherwise prior permission required
  • Results (output) may be used freely

Crossref: text-mining links for 100 % of recent and 60 % of older articles; licence information for 100 % and 60 % respectively.

KargerTDM permittedAI permitted

Key links

API key

  • no public API; TDM after prior request

Content, formats and quotas

  • Content as XML and PDF
  • FTP data service with weekly updates (aimed at industry customers)

Contract terms (excerpt)

  • Non-commercial, scholarly, project-based research, only during the term of the contract
  • TDM rules also apply to AI tools; training in a closed or self-hosted environment permitted for research
  • Bots and indexing only within the scope of TDM

Crossref: text-mining links for 51 % of recent and 78 % of older articles; licence information for 98 % and 93 % respectively.

Lippincott Williams & Wilkins (LWW)TDM conditionalAI not specified

Key links

  • no dedicated TDM page from the publisher

API key

  • no public API

Content, formats and quotas

  • No public information on delivery route, formats or quotas

Contract terms (excerpt)

  • Non-commercial, academic TDM, only for Wolters Kluwer Health content
  • No copies of substantial parts

Crossref (member Wolters Kluwer Health): text-mining links for 1 % of recent and 8 % of older articles; licence information for 33 % and 11 % respectively.

MDPITDM permittedAI permitted

Key links

API key

  • not required; FTP access on request

Content, formats and quotas

  • OAI-PMH: metadata in Dublin Core and AGRIS; sufficient for title, abstract, keywords and DOI
  • FTP bulk download: full texts of all articles (PDF, XML), updated twice a month
  • Crossref: ORCID, affiliations, funders and references require Crossref; MDPI deposits new metadata there within about 24 hours of publication (MDPI Blog).

Terms of use

  • The article's CC licence applies (usually CC BY): use with attribution

Crossref: hardly any text-mining links (0 % / 1 %); licence information for 100 % of recent and 98 % of older articles.

Oxford University Press (OUP)TDM conditionalAI not specified

Key links

API key

  • no public API; access beyond what is permitted only after whitelisting

Content, formats and quotas

  • No public information on formats or quotas

Contract terms (excerpt)

  • Non-commercial TDM by members of academic institutions
  • Copies only for TDM, not to be passed on, delete after use; attribution when publishing
  • No crawling

Crossref: text-mining links for 0 % of recent and 10 % of older articles; licence information for 99 % and 34 % respectively.

Public Library of Science (PLOS)TDM permittedAI permitted

Key links

API key

  • not required

Content, formats and quotas

  • Search API: metadata and article sections as XML or JSON; max. 100 rows per request; 7200 requests/day, 300/hour, 10/minute
  • allofPLOS: all full texts and metadata as JATS XML, approx. 5 GB, updated daily
  • Metadata with abstracts also via PubMed Central

Terms of use

  • No contract; all content under CC BY: use with attribution

Crossref: no text-mining links; licence information for 100 % of recent and 99 % of older articles.

Royal Society of Chemistry (RSC)TDM conditionalAI not specified

Key links

  • no public API documentation

API key

  • no public API; delivery route by agreement with RSC

Content, formats and quotas

  • TDM on locally loaded content; results on a university server

Contract terms (excerpt)

  • TDM permitted, delivery route by mutual agreement
  • No automated programs or web crawlers
  • Uses not mentioned, e.g. AI, require permission

Crossref: no text-mining links; licence information for 90 % of recent and 29 % of older articles.

SageTDM not specifiedAI not specifiedCrossref 99 % / 89 %

Key links

API key

  • no publisher API; retrieval via the Crossref links

Content, formats and quotas

  • Full texts via the text-mining links in the Crossref metadata
  • Weekdays 00:00–12:00 (Los Angeles time) max. 1 request every 6 seconds, otherwise 1 every 2 seconds

Terms of use

  • No valid contract with TDM provisions; only individual titles licensed
  • Website: non-commercial TDM of lawfully accessible content permitted; any AI use requires a licence from Sage

Crossref: text-mining links for 99 % of recent and 89 % of older articles; licence information for 99 % and 91 % respectively.

Springer NatureTDM conditionalAI conditionalCrossref 96 % / 89 %
Download tool: Data tool for Springer Nature, runs in the browser with a small local Python script, retrieval by DOI via the Meta, Open Access and TDM APIs.
  • Metadata with abstracts: CSV, JSON
  • Full texts: JATS XML (original), TXT

Key links

API key

Content, formats and quotas

  • TDM API: subscribed full texts, JATS XML; approx. 500 hits/day, 200/minute
  • Open Access API: over 1.5 million OA full texts, JATS XML; 500 hits/day, 100/minute
  • Meta API: metadata with abstracts for over 16 million documents; JSON, XML

Contract terms (excerpt)

  • Download, including automated, to an internal server
  • No generative AI with material from the TDM API
  • Deletion obligation; no systematic downloading outside the API

Crossref: text-mining links for 96 % of recent and 89 % of older articles; licence information for 96 % and 90 % respectively.

Taylor & FrancisTDM conditionalAI not specified

Key links

API key

  • no public API; access via a dedicated IP address

Content, formats and quotas

  • No public information on formats or quotas; delivery after project registration

Contract terms (excerpt)

  • Register every project in writing in advance
  • Academic research, non-commercial; results on the university's secure network
  • No robots outside the designated channels

Crossref: text-mining links for 30 % of recent and 4 % of older articles; licence information for 53 % and 8 % respectively.

University of California Press (UC Press)TDM not specifiedAI not specified

Key links

  • no public TDM or API documentation

API key

  • no public API

Terms of use

  • No package licence, individual titles only; no information on TDM

Crossref: hardly any text-mining links (0 % / 1 %); licence information for 13 % of recent and 2 % of older articles.

WileyTDM conditionalAI conditionalCrossref partly: 48 % / 94 %

Key links

API key

  • TDM token after accepting the click-through licence
  • Retrieval via the university's IP address

Content, formats and quotas

  • Wiley TDM API: full texts as PDF only, by DOI (api.wiley.com/onlinelibrary/tdm/v1/articles/{DOI})
  • Max. 3 requests per second

Contract terms (excerpt)

  • Access via the API only, no crawling
  • Loading onto own servers and indexing permitted
  • Deletion after the project ends

Crossref: text-mining links for 48 % of recent and 94 % of older articles, leading directly to the Wiley TDM API; licence information for 100 % and 99 % respectively.

Databases and metadata services

CrossrefTDM permitted (metadata)AI not specified
Tutorial: Crossref API, Jupyter notebook (Python) by Digital Scholarship, University Library Bern.
  • Works: metadata, queries, filters, facets, cursors, citation formats, retractions, relations
  • Funders, members, journals, prefixes, types and licences
  • Habanero Python client library

Key links

API key

  • not required; higher limits with the mailto parameter (“polite pool”)

Content, formats and quotas

  • REST API: metadata on works, journals, members, funders etc.; JSON
  • Per work, full-text links (field link, e.g. for text mining) and licence URLs (field license); filter full-text.application
  • Single-record requests 5 per second (10 with mailto), list requests 1 and 3 per second respectively

Terms of use

  • No contract; metadata “free to use”, some abstracts may be protected; funder data CC0
  • Access to the full text behind a Crossref link remains controlled by the publisher
DimensionsTDM conditionalAI not specified

Key links

API key

  • Analytics API key on request

Content, formats and quotas

  • Analytics API: queries via POST, response in JSON; token obtained from the API key, valid for about 2 hours; Python client Dimcli; max. 30 requests per minute, 1000 records per query, 50,000 per search
  • Analytics MCP: connects AI assistants (e.g. Claude Desktop, VS Code) using the API key; local or hosted

Contract terms (excerpt)

  • Copying, analysing and processing for internal purposes only, during the term; bulk data only on own, secured systems
  • No disclosure to third parties, no commercial use, no search functions or tools for third parties
  • Publication of data only with the consent of Digital Science and with attribution
  • AI is not covered by the contract; internal use only, not for third parties

Further terms per website

  • free access possible for eligible scientometric research (Scientometric research)
  • Analytics API not for local copies replacing the API and not for training purposes; limit use to the specific analysis task; Dimensions may block access in case of violations (DSL Usage Policy)
OpenAlexTDM permittedAI permitted
Tutorial: OpenAlex API, Jupyter notebook (Python) by Digital Scholarship, University Library Bern.
  • Works, authors, topics, institutions and sources endpoints
  • Search, filters, logical operators and grouping; large result sets
  • Results as a DataFrame (pandas)

Key links

API key

  • free key recommended (higher daily budget, cursor paging)

Content, formats and quotas

  • REST API: works, authors, sources, institutions, topics, publishers, funders; metadata with citation data and full-text links; JSON
  • Without key 10, with free key 30 requests per second; daily budget USD 0.10 and USD 1 respectively
  • Data Dump: free quarterly snapshot

Terms of use

  • No contract; data under CC0
PubMed Central (PMC)TDM permittedAI conditional

Key links

API key

  • not required

Content, formats and quotas

  • Open Access Subset: XML, TXT, PDF, BioC; split by licence (commercial use allowed, non-commercial only, other)
  • Author Manuscript Dataset: XML, TXT, BioC
  • XML via all access routes, TXT and PDF only via the Cloud Service (AWS); FTP discontinued

Terms of use

  • No contract; at least a TDM licence, otherwise the CC licence of the individual article
ScopusTDM not specifiedAI permitted

Key links

API key

  • Account on the Developer Portal and API key

Content, formats and quotas

  • Scopus Search API: metadata; JSON, Atom, XML; 20,000 requests/week, 9 per second
  • Abstract Retrieval API: abstracts linked to author and affiliation profiles; 10,000/week
  • Author Retrieval API: author profiles; 5000/week
  • Affiliation Retrieval API: institution profiles; JSON, XML; 5000/week
  • Serial Title API: journal data by ISSN with metrics (CiteScore, SJR, SNIP); JSON, XML; 20,000/week

Contract terms (excerpt)

  • TDM is neither permitted nor prohibited, i.e. not covered; analysing and processing only permitted with a closed AI tool
  • Access routes: robots, crawlers, automated downloads and systematic storage are prohibited; the contract does not cover the Scopus APIs
  • Uses not covered, such as TDM, can be notified to Elsevier; if there is no objection within 30 days, they are deemed permitted
  • AI use in a closed AI tool permitted, including training; open AI tools prohibited
  • Metadata on the university's own output may go into the repository
  • Per website, non-commercial use only