PLOS Section Literature Search Export

Search the PLOS corpus by title, abstract or article section using the source Solr phrase and Boolean syntax. Export a complete bounded publication-date selection as one DOI per full article for screening, corpus building and repeatable literature review. Data Provided by PLOS. This Actor is independently developed and is unaffiliated with PLOS.

Input

query: native PLOS Solr expression, default :, max 500 characters. Examples: title:malaria, abstract:"synthetic biology", materials_and_methods:CRISPR. publishedFrom and publishedTo: inclusive YYYY-MM-DD; omitted or empty bounds resolve to the seven-day UTC window ending yesterday. Historical windows up to 366 inclusive days are supported; future dates rejected. maxSafetyItems: 1–1000, default 1000. If the entire source match set cannot fit, the Actor fails refine-required before output or billing; narrow date/query. Exact section syntax matches PLOS, with no relevance inference.

Output

One type=research_article record per DOI: title, scholarly authors, abstract or null, publication timestamp, journal, articleType or null, subject paths, query, sourceUrl, sourceAttribution and retrievedAt. Always apply doc_type:full on every page to avoid the section-document DOI variants reported in community issues. Corrections and editorial article types retain their own DOI and type. Deterministic source ordering is publication_date desc,id asc; no citation or impact ranking. COVERAGE reports dates/query, sourceTotal, pages, selectedRows, complete selection and no truncation. Data Provided by PLOS appears in every row. Full-text bodies, PDF, affiliation/contact and altmetrics are excluded.

Limits and failures

Lightweight official API JSON, 256 MB; no key, proxy or paid service. At most 12 GET attempts including robots; <=100 rows/request, sequential >=6 second spacing (<=10/minute); 4 MB wire/decoded per response, 64 MB total transfer, 24 MB staged output and 240 seconds acquisition. Maximum 10 data pages/1000 articles. Validate total, start, cardinality, full DOI, required fields, dates and duplicate IDs before any delivery. Moving source counts, missing fields, section-ID leakage, invalid query responses, robots refusal, auth blocks, 429, timeouts and malformed JSON fail explicitly. Zero source matches fail no-matching-results with uncharged COVERAGE. Delivery has a separate 90-second deadline; the Actor cloud timeout setting is 360 seconds, covering the 240-second acquisition plus 90-second delivery phases with startup margin. Frequent schedules across runs must still respect PLOS hourly/daily API quotas and the shared-IP limit.

Privacy and rights

Only scholarly author display names and article metadata. Omit contact/affiliation/corresponding-author fields, emails/phones, raw responses and account state. Only https://api.plos.org/search is requested after its origin robots check. PLOS API display policy requires attribution; every row says Data Provided by PLOS. Source text and research findings must be cited and interpreted in context. PLOS permits mining, reuse and sharing for any purpose; see https://plos.org/text-and-data-mining/. No PLOS logos or implied sponsorship.

Pricing

Published price: $0.001 per unique DOI actually delivered using the record-delivered primary event. Private cloud probes verified 435/435 default, 8/8 alternate and 529/529 maximum-load dataset rows to primary events. No custom start charge or actor-defined free rows; platform compute may apply. Source scanning, metadata summaries and duplicate handling are uncharged. PPE allowance must cover the complete result before delivery, rechecked for each batch with exact charged-count parity. Non-PPE local runs save data and summaries without charged events. Private cloud runs confirmed DC access and primary event counts on the reviewed build.