Sources & provenance

Every source we use, in the open.

Tensorust is built only on public-domain or openly licensed records, accessed through official APIs and bulk downloads. No scraping, no paywalled data, no private information. Here is every dataset, its license, and how we get it, so you can check the supply chain behind any number on this site.

✓ Public-domain or openly licensed ✓ Official APIs, no scraping ✓ Source + hash kept on every record
SourceWhat it coversLicenseHow we access itRefresh
USAspending.gov
FPDS		Federal contract & grant awardsU.S. public domainOfficial APILive on query
SEC EDGARPublic-company filings & XBRL financialsU.S. public domainOfficial APILive / daily
U.S. Senate LDAFederal lobbying disclosuresU.S. public domainOfficial APIQuarterly
NIH RePORTERNIH research grants & projectsU.S. public domainOfficial APILive on query
NSF AwardsNational Science Foundation grantsU.S. public domainOfficial APIPeriodic
openFDA
Drugs@FDA		Approved drugs & sponsorsU.S. public domainOfficial APIWeekly
Federal RegisterRules, notices, executive ordersU.S. public domainOfficial APIDaily
CISA KEVActively-exploited vulnerabilitiesU.S. public domainOfficial JSON feedLive
NVD (NIST)Reported vulnerabilities (CVEs)U.S. public domainOfficial APIPeriodic
GLEIFLegal-entity identifiers (LEI)CC0Official APIPeriodic
PubMed / NLMBiomedical literatureU.S. public domainOfficial E-utilities APILive on query
OpenAlexScholarly works, all fieldsCC0Official API / snapshotPeriodic
CrossrefScholarly metadata & DOIsOpen (CC0 facts)Official APIPeriodic
arXivPhysics / CS / AI preprintsFree reuse (metadata)Official APIPeriodic
How provenance is kept. Every record we collect is stored with the exact source URL it came from and a sha256 hash of the original payload, in a tamper-evident log. That is why any figure on this site can be traced back to the government or publisher response it was parsed from. If you ever find a number you cannot verify at its source, email mattbusel@gmail.com and it will be corrected.
Your data stays yours. These are the only inputs we use. We store no private data, and nothing you type into a tool is sold, shared, or used to train anything.