Discovery metadata
Titles, authors, subjects, language, DOI, ISBN, repository, landing pages, and candidate PDF links.
Scholarly data product / dated and reproducible
Find scholarly PDFs with rights evidence attached.
A metadata-only catalog for scientific search, RAG, research assistants, repository monitoring, and rights-aware corpus acquisition. Every record preserves its source, operational rights basis, and link evidence.
Metadata and links only. No PDFs or extracted full text are included.
10,000free evaluation rows
8,500freshly checked ready links
37normalized evidence fields
0PDF files redistributed
What is included
A production-shaped acquisition index—not a pile of URLs.
Use normalized fields and evidence to decide what to retrieve, exclude, review, or refresh before content enters a downstream system.
Rights evidence
Raw and normalized license fields, evidence basis, attribution completeness, NASA determinations, and conservative review tiers.
Link evidence
Recorded URL status, validation timestamps, redirects, response metadata, PDF signature results, and retryable failure classes.
Reproducible delivery
Schema-identical Parquet shards, exact row counts, SHA-256 manifests, deterministic samples, and machine-readable QA reports.
Validation method
Evidence that says exactly what it proves.
Between August 31 and September 2, 2026, two deterministic workers checked all 776,075 distinct non-empty PDF URLs. A streamed ranged GET read no more than 16 KiB, followed redirects, and required a PDF signature within the first 1,024 bytes. 611,609 URLs (78.81%) passed. This verifies that a URL began serving PDF content at the recorded time; it does not claim full-file integrity, document identity, or legal permission to reuse a linked work.
- 1Normalize and deduplicate
Preserve source IDs while flagging DOI overlap and source-specific anomalies.
- 2Classify rights evidence
Keep Creative Commons evidence, NASA public-use determinations, review cases, and blocked cases distinct.
- 3Probe and record
Store status, redirect, HTTP metadata, timestamps, PDF version, and failure evidence without retaining document bodies.
- 4Validate the release
Check schemas, counts, source-ID uniqueness, tier invariants, manifests, and checksums.
Rights boundary
Conservative tiers for operational decisions.
A
Recognized commercial Creative Commons or CC0 URL; license obligations still apply.
N
Official NASA public-use evidence without captured risk flags; never relabeled CC-BY.
B
Plausible rights assertion without sufficiently strong machine-checkable evidence.
C
Manual review or authoritative evidence resolution required.
D
Restrictive or adverse evidence; do not infer redistribution permission.
Operational classifications are not legal opinions and do not grant rights to linked documents. Linked works remain governed by their respective rights holders and terms.
Launch pricing
Evaluate free. License the complete snapshot when it fits.
Evaluation
$0
10,000-row validated sample, full schema, rights methodology, checksums, and interactive explorer.
Download the sampleFounding commercial snapshot
$499 USD
The standard commercial snapshot license at an introductory price for the first ten customers. One legal entity, up to five authorized users.
Buy securely with PayPalStandard / custom
$1,250 USD
Standard dated snapshot. Team expansion, fresh scoped validation, filtered exports, and customer-cloud delivery start at $2,500.
View gated listingA snapshot purchase covers the identified dated release. Recurring updates and service-level commitments are not included unless stated in a separate order or invoice.
Start with evidence
Search the sample before you buy.
Inspect real records, rights tiers, link outcomes, and code-ready Parquet delivery.