THE RECORD

← Back to List

← Back to Essays

OPEN THE RECORD COMPANION AI →

Opens in a new window. Powered by a custom GPT trained on The Record's complete master document.

1. Purpose and Scope

The Record is a chronological accountability archive documenting the political career of Donald Trump from pre-political history through the present day. It is not a news site, a blog, or an opinion column. It is a structured dataset of verifiable events, organized for pattern recognition across time. The archive spans from 1927 — when Fred Trump was arrested at a Ku Klux Klan rally in Queens — through the current second presidential term. Every entry is designed to answer a simple question: what actually happened, when, and why does it matter?

2. The Three-Layer Entry Format

Every entry in The Record uses a strict three-layer format designed to separate fact from analysis from editorial interpretation. THE FACTS is the first layer — written in neutral, journalistic prose describing what happened, sourced to verifiable reporting. This layer reads like wire copy: no adjectives of judgment, no "clearly" or "shockingly," no opinion. SIGNIFICANCE is the analytical layer, explaining why the event matters in context — what pattern it fits, what precedent it sets, what power dynamic it reveals. GOALPOST is the most editorial layer, documenting the rhetorical shift or norm erosion — the actual talking points used to defend or normalize the event, and how those defenses compare to earlier stated positions. These three layers are always visible and always labeled, so readers can distinguish sourced facts from informed analysis from pointed editorial observation.

3. Sourcing Standards

Every entry in The Record is backed by at least one source citation, and the archive maintains a zero-tolerance policy for unsourced claims in the fact layer. Sources are drawn from major wire services (AP, Reuters), national newspapers of record (New York Times, Washington Post), broadcast journalism (CNN, NBC, CBS, ABC, PBS, BBC, NPR), government records (Congress.gov, White House archives, FEC filings, court documents via CourtListener), and specialized investigative outlets. The archive currently contains over 5,300 source citations across 180 unique domains. Wikipedia and Google search fallback links have been systematically replaced with primary reporting. All source URLs use HTTPS.

4. Entry Types and Date Precision

The Record distinguishes between two types of entries. Event entries document a specific occurrence on a specific date — a speech, a firing, a court ruling, a tweet. These make up the vast majority of the archive. Context entries, marked with a purple border, document broader patterns, systemic developments, or periods that span weeks or months and cannot be reduced to a single date. Each entry also carries a date precision field: "day" for events with a confirmed specific date, "month" for events known only to have occurred during a particular month, and "year" for historical events where only the year is documented. This precision metadata is displayed to the reader so they know exactly how confident the dating is.

5. The Six Eras

The archive is organized into six chronological eras that correspond to distinct phases of Trump's political trajectory. Formation covers everything before June 16, 2015 — the family history, business career, early political flirtations, and the psychological and financial architecture that produced the candidacy. Campaign 1 runs from the golden escalator announcement through Inauguration Day 2017. Term 1 covers the entire first presidency. Post-Presidency spans from January 20, 2021 through November 14, 2022 — the January 6 aftermath, legal exposure, and the beginning of political rehabilitation. Campaign 2 covers the second campaign through Inauguration Day 2025. Term 2 covers the current second presidency. Each era has its own companion document with every entry from that period.

6. Tagging and Classification

The Record uses 30 automated topic tags powered by regex pattern matching against entry text. Tags include Russia/Mueller, Immigration, Courts & Legal, Economy & Trade, COVID-19, January 6, Elections, Ukraine, Media & Press, DOGE & Spending, Epstein, Culture War, Deaths & Consequences, Follow the Money, and others. Entries can match multiple tags simultaneously. In addition to topic tags, entries are classified as "Major" (significant enough to be highlighted in filtered views) or "Goalpost" (documenting a specific instance of norm erosion, rhetorical shifting, or moved goalposts). These classifications are editorial judgments made by the editor and are subject to revision.

7. AI Transparency

The Record is built with two AI systems: Claude (by Anthropic) and ChatGPT 5.4 Extended Thinking (by OpenAI). AI is used for drafting entry text from source material, analytical assistance in the significance and goalpost layers, code generation for the HTML app and document generators, and quality assurance auditing. The editor reviews, edits, and approves all AI-generated content before it enters the archive. AI-generated opinion essays (the "AI Opinion" section) are clearly separated from the factual timeline and explicitly labeled as AI-generated analysis. The archive does not use AI to fabricate events, invent sources, or generate unverified claims. Every factual assertion in the FACTS layer is traceable to a cited source.

8. The Pause Before Harm Protocol

This project operates under the Pause Before Harm Protocol (PBHP), an open-source harm-reduction framework for AI and human decision-making. PBHP asks: "If I'm wrong, who pays first — and can they recover?" It prescribes a 7-step decision flow: name the action, identify constraints, check for impermissible constraints, identify harms to the least-powerful stakeholders first, rate each harm, assign a risk gate, and act on the gate with logging. Applied to The Record, PBHP means: every entry must name who is harmed by the documented events, language must maintain "brutal clarity with zero contempt" (no mockery, no dehumanization), and the archive must be transparent about its own methods, limitations, and potential biases. The protocol is available at github.com/PauseBeforeHarmProtocol/pbhp.

9. Companion Documents and Master PDF

The Record produces 34 companion DOCX documents that provide focused views into the data: 6 era documents (one per era), 16 topic documents (Courts, Democracy, Media, January 6, Epstein, Foreign Influence, etc.), 5 AI opinion essays, and documents covering methodology, editorial approach, politicians, people mentioned, and internal audits. These documents are regenerated from the same source data as the HTML app, ensuring consistency. All 34 documents are compiled into a single master PDF (THE-RECORD-COMPLETE.pdf) that runs approximately 2,900 pages. Every companion document includes full attribution, contact information, and project links.

10. Corrections, Limitations, and Accountability

The Record is maintained by a single editor, Phillip Linstrum, a QA professional. This is both a strength and a limitation — it ensures consistent editorial standards but means the archive reflects one person's judgment about what events to include and how to frame their significance. The editor is personally accountable for every entry and welcomes corrections. If you identify a factual error, a missing source, a misattribution, or an unfair characterization in any layer, report it to pausebeforeharmprotocol_pbhp@protonmail.com or via facebook.com/plinst. Corrections are made transparently. The goal is accuracy and completeness, not advocacy — though the archive makes no pretense of neutrality about whether accountability matters. It does. That is the premise. Everything else is evidence.

What This Is

The Record is a chronological, factual archive of Donald Trump's political career — from his father Fred's KKK arrest in 1927 through the present day. Every entry uses a three-layer format: THE FACTS (what happened), SIGNIFICANCE (why it matters in context), and GOALPOST (the argument used to normalize it). The archive does not editorialize in the fact layer. It does not hide its analysis. It labels everything.

Entry Types

Specific Event entries document a particular event on a particular date with verifiable facts. Context entries (marked with a purple border) document broader patterns, periods, or systemic developments that span time ranges. Context entries are labeled to distinguish them from single-day events.

Editor & Accountability

Maintained by Phillip Linstrum, a QA professional. Built with Claude (Anthropic) and ChatGPT 5.4 Extended Thinking (OpenAI). This project is self-funded and independent. The editor is personally accountable for every entry. AI-generated content (the "AI Opinion" section) is clearly labeled as such and separated from the factual timeline.

Support This Project

Methodology

Facts are drawn from major news reporting (AP, Reuters, NYT, Washington Post, CNN, BBC, court filings, congressional records, and official government documents). The significance layer provides analytical context grounded in documented evidence. The goalpost layer documents the actual rhetorical defenses used by supporters — not as straw men, but as observed talking points. Entries are tagged by topic and era, and filtered by automated keyword matching.

Corrections Policy

If you identify a factual error, missing context, or misattribution, please report it. Corrections will be made transparently with timestamps. The goal is accuracy, not advocacy. Contact: pausebeforeharmprotocol_pbhp@protonmail.com · facebook.com/plinst

Pause Before Harm Protocol

This project is built through the Pause Before Harm Protocol (PBHP) framework — an approach to responsible AI use that prioritizes transparency, evidence, and accountability over speed or persuasion.

Companion AI

Have questions about The Record? Ask The Record Companion AI — a custom GPT trained on the complete master document. It can answer questions about any entry, era, topic, sourcing, methodology, or the Pause Before Harm Protocol.

Data & Technical

Last updated: May 2026. Source data maintained in structured JavaScript format with automated build pipeline. 34 companion DOCX documents provide deep-dive analysis by topic.