The UAP Index: Source-Based Data Capture and Trust Methodology

How the UAP Index converts source narratives into structured, searchable, human-reviewed records without treating claims as established facts.

Draft v1.0 – Website Reference Document

Core operating rule: The UAP Index stores claims as attributed claims, not as established facts. It records what a source says, who said it, how they claim to know it, and what other records may support or challenge it.

1. Purpose of the UAP Index

The UAP Index is a source-based research database. Its purpose is to preserve source material and organize the claims, events, people, evidence references, corroborations, contradictions, and topics contained in that material so that users can search and analyze the record in a disciplined way.

The UAP Index is not designed to prove what UAPs are. It is not a truth engine. It is a structured index of what sources say, who said it, where it came from, and what other records appear to support, contradict, or contextualize those statements.

The central purpose is to make UAP information easier to inspect, compare, filter, and analyze without losing the original source context.

2. Core Principle: Claims Are Stored as Claims, Not Facts

The most important rule is that the database stores attributed statements rather than conclusions. A statement may later be supported, challenged, refined, or rejected, but the database does not convert that statement into a fact at the capture stage.

Incorrect: The Tic Tac accelerated rapidly.

Correct: David Fravor stated that the Tic Tac object rapidly accelerated and disappeared.

This distinction keeps the database neutral, defensible, and useful. The UAP Index preserves what was said and where it was said. It does not declare that every statement is true.

3. What the Database Does With a Source Narrative

Most UAP information arrives as a narrative: a YouTube interview, podcast, article, report, hearing transcript, PDF, book excerpt, witness account, or documentary segment. The UAP Index decomposes that narrative into structured records.

A single source may contain many different pieces of information: named people, organizations, dates, events, claims, observations, evidence references, contradictions, and themes. The database separates those elements so they can be searched and analyzed later.

The system does not interpret the narrative as true or false. It indexes the narrative into parameters and categories while preserving attribution and uncertainty.

4. Primary Data Elements Captured

Source: the original material, such as a video, podcast, article, report, hearing, PDF, interview, or witness submission.

Transcript or source text: the preserved text of the source, stored locally when available so the record survives if the original source is removed or changed.

Person: a witness, speaker, interviewer, official, analyst, author, researcher, public figure, or other named individual.

Organization: an agency, military unit, media outlet, nonprofit, program, contractor, office, publication, or group connected to the source or event.

Event: the UAP incident, hearing, interview, article, report, disclosure milestone, or other distinct occurrence being discussed.

Statement: an attributed claim, observation, assessment, opinion, correction, report, or testimony extracted from the source.

Evidence reference: a referenced video, radar track, photograph, document, sensor record, witness testimony, physical material, or other evidence type.

Corroboration: another source, statement, witness, document, or evidence item that appears to support a statement or part of a statement.

Contradiction: another source, statement, witness, document, or evidence item that appears to dispute, correct, narrow, or conflict with a statement.

Tags/topics/categories: searchable labels applied to sources and other entities so users can find related material.

5. Source Intake and Preservation

Every source begins with a source record. The source record captures basic provenance: title, URL, source type, publisher, author or host, publication date if known, access date, and a neutral summary.

When possible, the full transcript or text is archived in the database. This is important because videos can be deleted, transcripts can disappear, articles can be edited, and links can break. The archived text becomes the preserved evidence trail for the extraction.

The archived transcript is not treated as proof that the claims are true. It is treated as proof of what the source contained at the time it was captured.

6. AI Extraction and Human Review

AI may be used to extract candidate records from source text, but AI output does not go directly into final production tables. Extracted items first enter a human-readable review queue.

The review queue allows a human reviewer to approve, reject, or mark items as needing edits. This protects the database from hallucinated names, duplicate entities, wrong dates, inferred identities, weak labels, and claims accidentally rewritten as facts.

Only reviewed and approved records are inserted into final database tables.

7. Review Status Definitions

Pending: the extracted item has not yet been reviewed.

Approved: the item is acceptable for insertion into final tables.

Needs Edit: the item is useful but requires correction, verification, or clarification before final insertion.

Rejected: the item should not be inserted into final tables.

This process is designed to create a controlled path from raw source material to public search results.

8. Event Separation Rule

The original UAP event must be separated from later interviews, articles, hearings, reports, or commentary about that event.

For example, the 2004 Nimitz Tic Tac encounter is an event. A 2023 podcast interview with David Fravor is a separate source event. A 2017 New York Times article discussing the case is another separate publication/disclosure event.

Separating these prevents timelines from becoming confused and keeps the database from treating later discussion as the original incident.

9. Statement Capture Rules

Statements should be atomic. A long claim containing multiple ideas should be broken into smaller statement records when possible.

Each statement should identify the speaker, the source, the related event, the statement type, and the knowledge basis.

Statement types may include Observation, Testimony, Assessment, Analysis, Opinion, Hypothesis, Report, Correction, or Conclusion.

Knowledge basis describes how the speaker claims to know the information: first-hand observation, direct participation, second-hand testimony, document review, evidence review, analytical assessment, or unknown.

10. Corroboration Handling

Corroboration is recorded when another source, statement, witness, document, sensor record, or evidence item appears to support the same statement or part of the same statement.

Corroboration does not mean proof. It means the database has identified another record that appears to support or align with the statement.

For example, if one pilot states that a radar contact existed and another source references shipboard radar observations, the system may record a corroboration relationship, while still preserving the original attribution and uncertainty.

Corroboration should include a type, a supporting record, a description, a status, and a strength estimate when appropriate. Strength should be conservative: weak, moderate, strong, or unknown.

11. Contradiction Handling

Contradictions are recorded when another source, statement, witness, document, or evidence item appears to dispute, correct, narrow, or conflict with a statement.

A contradiction does not automatically make the original statement false. It simply marks the record as contested or corrected by another record.

For example, if one claim says a video was ten minutes long and another source states the video was approximately one minute and thirty seconds, the database should preserve both the disputed claim and the correcting statement.

12. Source Categorization and Topic Tagging

Source categorization is essential. Users should be able to find sources by topic, not just by title. A user may want to see every source that discusses AARO, USO cases, Nimitz, crash retrieval claims, radar evidence, whistleblowers, or pilot testimony.

Therefore each source should be tagged at the source level. Source tags describe what the source discusses, not what the source proves.

Example: tagging a source with Nimitz Tic Tac means the source discusses the Nimitz Tic Tac case. It does not mean the database has concluded that every claim in that source is true.

Tags should also be available for other entities, such as events, people, statements, organizations, and evidence items. This allows both broad source discovery and precise statement-level analysis.

13. Recommended Topic Categories

People: David Fravor, Ryan Graves, David Grusch, Bob Lazar, Luis Elizondo, Jay Stratton, and other individuals.

Organizations: U.S. Navy, AARO, AATIP, NASA, The New York Times, Americans for Safe Aerospace, military units, agencies, and programs.

Events: 2004 Nimitz Tic Tac Encounter, 2023 Congressional UAP Hearing, 2017 New York Times UAP Article, and other distinct cases or milestones.

Evidence Types: radar, FLIR, video, photograph, witness testimony, official document, physical material, sensor data, and other evidence classes.

Observation Characteristics: rapid acceleration, hovering, no visible propulsion, no wings, no rotor wash, transmedium movement, high-speed descent, instant direction change, low observability.

Claim Categories: crash retrieval, reverse engineering, non-human intelligence, military encounter, aviation safety risk, government secrecy, whistleblower testimony, classified program.

Domains: airborne, underwater/USO, space, ground, transmedium.

Policy and Disclosure: congressional hearings, AARO, reporting systems, whistleblower protections, declassification, public transparency.

14. Source-Level Versus Statement-Level Tags

Source-level tags describe the overall content of a source. For example, a long interview may be tagged with Nimitz Tic Tac, David Fravor, AATIP, UAP Disclosure, Aviation Safety, and Pilot Testimony.

Statement-level tags describe the content of a specific statement. For example, one statement may be tagged with rapid acceleration; another with no rotor wash; another with FLIR video length; another with no NDA.

Using both levels allows users to search broadly for sources and narrowly for specific claims or characteristics.

15. Public Website Use

The public website should not expose raw database tables directly. Public users should interact with read-only views designed for safe searching and browsing.

Public views can show event summaries, source indexes, statement search results, evidence summaries, and topic pages. They should not expose internal review notes, import logs, raw pending records, private metadata, or administrative fields.

Authenticated users may eventually submit UAP accounts or source suggestions, but those submissions should enter a separate staging table. Public users should never be allowed to write directly into final production tables.

16. Trust Through Process

The trust value of the UAP Index comes from process discipline: preserve sources, separate claims from facts, require attribution, document uncertainty, stage AI extraction for human review, and record corroboration and contradiction without overstating conclusions.

The database becomes more valuable as it grows because it allows users to compare sources across events, people, evidence types, topics, and knowledge basis. Users can ask better questions because the data has been organized without stripping away context.

The UAP Index does not ask users to trust a conclusion. It gives users a structured way to inspect the record themselves.

17. Summary Statement

The UAP Index is a source-based research database that preserves original source material and decomposes each narrative into structured, attributed records. It identifies people, organizations, events, statements, evidence references, corroborations, contradictions, and searchable topics without deciding whether the claims are true. Each extracted item is staged for human review before being inserted into final database tables. The system allows users to search and compare approved records by source, speaker, event, evidence type, knowledge basis, topic, and relationship.

In plain language: the UAP Index carefully cuts source narratives into searchable pieces while preserving who said what, where it came from, and what other records support or challenge it.

18. Example: How One Source Becomes Structured Data

A single source narrative can produce multiple database records. The table below illustrates the concept.

Narrative element	Database treatment	Neutral handling rule
David Fravor is interviewed	Person + source speaker/subject	Identify the person; do not evaluate truth.
Discussion of the 2004 Tic Tac event	Event record	Separate original event from later interview.
Fravor says he saw a Tic Tac object	Attributed statement	Store as Fravor stated, not as confirmed fact.
FLIR video is referenced	Evidence reference	Record that a video was referenced and link it to source/event.
Another source supports a point	Corroboration relationship	Record support without declaring proof.
Another source disputes a point	Contradiction relationship	Record conflict without deciding final truth.
AATIP, AARO, Nimitz are discussed	Source/topic tags	Tag as discussed topics for search and discovery.

19. Example Source Tags for a Fravor/Nimitz Interview

Topic	Category	Relevance	Meaning
Nimitz Tic Tac	Case/Event	High	The source substantially discusses the Nimitz encounter.
David Fravor	Person	High	The source features or discusses David Fravor.
Pilot Testimony	Evidence Type	High	The source contains pilot testimony or pilot-based accounts.
AATIP	Program	Medium	The source discusses AATIP in context.
AARO	Program	Medium	The source discusses AARO or reporting pathways.
Aviation Safety	Theme	Medium	The source discusses operational safety concerns.
Radar Evidence	Evidence Type	Medium	The source references radar/sensor observations.
UAP Disclosure	Policy/Disclosure	Medium	The source discusses public disclosure, government response, or stigma.

Prepared for The UAP Index / UAP Signal Report as a public trust and methodology reference.