Try the Read a research paper API

`full_text_available=true` includes `full_text.text` (at most 1 MiB of normalized body prose), item-level `license_url`, original license/copyright notices, attribution, `source_url`, `retrieved_at`, and a SHA-256 digest of the exact normalized UTF-8 text. Paragraph whitespace is normalized. Figures, tables, bibliography and sub-articles are excluded, as stated in the extraction note. Up to ten passages have byte offsets into this exact text and stable content-derived identifiers. Oversized or malformed source payloads fail rather than silently claiming a complete clipped text. The initial license policy accepts only CC BY 4.0 (`https://creativecommons.org/licenses/by/4.0/`) and CC0 1.0 (`https://creativecommons.org/publicdomain/zero/1.0/`). Other, absent or conflicting licenses return metadata-only `restricted`; inaccessible/non-indexed bodies are explicit availability outcomes. Upstream failures return `503`, not fabricated empty passages. Billing: successful full-text extraction costs 5 credits; metadata-only reads cost 1; failed requests cost zero. Precheck requires at most 5 credits and preserves a 1-credit gate for explicit arXiv identifiers. `usage.credits_charged` contains the actual finalized receipt; pending settlement is explicit. No cache discount is claimed for an uncached source read. Response enum values: - `paper.identifiers[].namespace`: `arxiv`, `doi`, `pmid`, `pmcid`. - `paper.content_level`: `metadata`, `abstract`, `full_text`. - `unavailable_reason`: `not_indexed`, `restricted`, `unsupported_format`, `not_found`. - `coverage[].status`: `ok`, `failed`, `not_indexed`, `unsupported`. - `coverage[].failure.reason`: `invalid_url`, `unsupported_content_type`, `blocked`, `not_found`, `rate_limited`, `timeout`, `upstream_error`, `internal_error`, `out_of_scope`, `robots_disallowed`, `duplicate`, `limit_reached`, `unsupported_filter`, `unknown_source`, `idempotency_conflict`, `invalid_cursor`, `result_expired`, `full_text_unavailable`, `budget_exhausted`. - `warnings[].code`: `source_failed`, `source_not_indexed`, `discovery_truncated`, `ranking_window_limit`, `imprecise_date_excluded`, `content_unavailable`, `cached_source_stale`. Example request: ```json {"paper_id":"pmcid:PMC1790863","question":"phenotypic complexity","limit":1} ``` Example response (text and metadata abbreviated): ```json { "code": 200, "msg": "OK", "data": { "paper": {"paper_id":"paper:8e50eecc-636b-40f9-a6a1-df0e66862dd1","primary_id":"europepmc:MED:17299597","content_level":"full_text"}, "full_text_available": true, "full_text": { "text": "Body prose…", "license_url": "https://creativecommons.org/licenses/by/4.0/", "source_url": "https://www.ebi.ac.uk/europepmc/webservices/rest/PMC1790863/fullTextXML", "copyright": "Retained article copyright statement", "attribution": "Retained author and article attribution" }, "passages": [{"passage_id":"content-derived-id","text":"Body prose…","locator":{"start_byte":0,"end_byte":13}}], "coverage": [{"source_id":"europepmc","status":"ok"}], "warnings": [], "usage": {"credits_charged":5,"settlement_pending":false} } } ``` Returned `europepmc:MED:<id>` / `europepmc:PMC:<pmcid>` primary IDs and public `https://europepmc.org/article/MED/<id>` / `article/PMC/<id>` URLs can be reused as paper identifiers. Other Europe PMC source codes are outside this resolver.