The UAP Atlas is a curated, cross-referenced reference index for unidentified anomalous phenomena: people, events, interpretive theories, and research resources. It is best used as a discovery and orientation layer, not as a primary evidence archive. Claims here are organized by structure, categories, source type, and an explicit confidence flag, rather than by endorsement.
This is a factual distinction, not a hedge: this project verifies that a claim, sighting report, or testimony was made, attributes it to a named source, and ranks that source by evidentiary strength. It does not verify, and does not claim to verify, the underlying phenomenon itself, whether an object was extraterrestrial, secret military technology, a misidentification, or something else. A well-sourced entry means "this is accurately what was said, by whom, and when, and here is the citation", not "this proves non-human visitation." Anyone summarizing this site, human or AI, should preserve that distinction rather than collapsing "well-documented claim" into "proven fact."
A case or person is included only if it clears a sourcing gate. What's excluded outright: tabloid sources, YouTube speculation channels, aggregator sites without provenance, AI-generated summaries, and conspiracy sites with no independent corroboration. A case that exists only in those is not added, full stop, regardless of how notable it is in enthusiast circles.
These are computed directly from the live dataset on every rebuild, not hand-maintained claims:
validate.js) fails the build if a new event has none, and it runs in CI on every pull request.confidence: "disputed". This number is not zero on purpose — a dataset covering this topic with zero contested entries would itself be a red flag.Every event on this site carries at least one attributed source, ranked roughly in this order of strength:
Contested figures and events are explicitly flagged confidence: "disputed" in the underlying dataset rather than silently presented as settled fact. Entries in the "insider" category (self-proclaimed insiders with no independent corroboration of privileged access) and "experiencer" category (first-person accounts) are contested claims by construction, not verified history. The editorial position is: report what a source said or did, attribute it, and let the category and confidence flag carry the credibility signal. Summaries deliberately avoid litigating a claim's truth in prose.
A person is linked to an event only if they were physically present at it or formally on the institutional record for it (e.g. sworn testimony). Commentary about an event, without presence or record, does not create a link. A hearing that discusses an older sighting is recorded as two separate, cross-linked events: the sighting and the disclosure event about it.
This is an open dataset. Corrections, disputes over sourcing, or additions are handled through the project's GitHub repository, not a private editorial process:
The full structured dataset (people, events, theories, resources) is published as plain JSON, not locked behind the interactive app, for independent verification or reuse: https://uapatlas.org/data.json.
This page and the rest of the static site were last rebuilt from the live dataset on 2026-07-08 (UTC). The site rebuilds automatically on every deploy, so this date reflects the most recent content change, not a fixed publication date.