Corpus Manifest

Every data source feeding Hermes is listed here with its role, approximate row count, license, and snapshot date. Nothing analyzed by Hermes comes from an un-disclosed source.

Totals

Manifest generated (UTC)2026-06-09T03:33:25.055328Z
Methodology version0.7.0
Total index rows87463
Date range1906-11-11 2026-04-22
Integrity hash (sha256 of sorted case_ids)A0271D6814576FB2
HERMES: 5 NUFORC: 87458

Data sources

IDNUFORC-archive
NameNational UFO Reporting Center (NUFORC) public archive
Rolehistorical corpus for statistical analysis
Rows87458
Snapshot date2019-09-17 (truthiswill fork); imported 2026-04-23
LicensePublic reports; redistributed via planetsig/ufo-reports -> truthiswill/ufo-reports fork CSV mirror
Reports are first-person witness accounts. Quality is uneven. Hermes treats NUFORC as a reporting-behavior corpus, not a phenomenon corpus. v0.7.0 upgraded from ufo-scrubbed CSV (79,636 rows, max 2014-05-08) to ufo-complete CSV (87,458 rows retained after lat/lon validation, same date ceiling) for improved pre-2014 coverage. Gap 2014-05-09 -> 2026-04-04 remains.
IDhermes-native
NameHermes native submissions
Rolelive witness submissions with contemporaneous analysis
Rows12
Snapshot datelive
LicenseSubmitted under Hermes Terms of Service (see /terms)
URL/
Each native case carries the full audit trail (see /docs/rules) at submission time.
IDephem
NamePyEphem astronomical computation
Rolemoon/sun/planet position per sighting
Rowsper-case
Snapshot datecomputed at case submission
LicenseLGPL (PyEphem library)
Local computation; no network calls for celestial state.
IDvisual-crossing
NameVisual Crossing historical weather
Rolehistorical weather at sighting time and location
Rowsper-case (on demand)
Snapshot dateper-case
Licensecommercial API key; data retained per-case
Weather values retrieved at submission time are stored verbatim in the case record for audit.
IDn2yo
NameN2YO satellite catalog (live snapshot)
Rolecatalogued objects overhead at intake time
Rowsper-case snapshot
Snapshot dateper-case
Licensecommercial API key; snapshot retained per-case
Intake-time snapshot only. Historical TLE propagation for exact-time line-of-sight is planned (SAT-LOS-01).
IDcelestrak-tle
NameCelesTrak general-perturbations TLE feed
Rolesatellite ephemerides for SAT-LOS-01 line-of-sight elimination
Rowsstations + visual catalog (~170 objects); refreshed daily
Snapshot datecached daily at /app/cache/tle/
LicenseCelesTrak redistribution policy (public, attribution requested)
sgp4 propagation performed locally; no data sent to CelesTrak beyond the daily GET.
IDskylink-adsb
NameSkylink ADS-B aircraft snapshot
Roleaircraft in 50 nm radius at intake time
Rowsper-case snapshot
Snapshot dateper-case
Licensecommercial API key; snapshot retained per-case
Intake-time snapshot only. Historical ADS-B backfill for exact sighting time is planned (AC-HIST-01).

Machine-readable version: /api/corpus. When a data source changes, the integrity hash changes, and the methodology changelog records why.