Methodology

01 // Primary sourcing

Every broadcast is built from primary source documents, raw data, archived recordings, and unaltered metadata. Secondary commentary is flagged as commentary. Where a story depends on a single document — a flight log, an unsealed deposition, a geophysical timeline — that document is named explicitly. If it can be linked legally, it is.

Citation links from the broadcast description are surfaced on each dispatch page in the Sources & receipts block. Self-links (channel, store, social) and affiliate URLs are stripped before display so the section stays a credibility signal, not a promotional list.

02 // Speculation framing

Where a claim is speculative, the broadcast says so explicitly: "the host argues", "according to the source", "the document states". Dispatch pages on this site preserve that framing word-for-word. Nexor will never assert a contested claim as the site's own statement of fact, and the LLM-cleaned transcript follows the same rule. If the host hedged, the page hedges.

03 // Algorithmic neutrality

The investigation bypasses algorithmic curation. The feed is not trusted. The signal is traced to its origin. Topics are chosen by editorial judgement, not by what the recommendation engine surfaces. This is the only filter that survives long enough to matter.

04 // Corrections policy

Errors are corrected in-thread on the original broadcast description and, where material, called out at the top of the dispatch page. Past dispatches are never deleted to bury a mistake. Retraction-by-deletion is treated as misconduct.

05 // Dispatch page provenance

Each dispatch page is built from the matching YouTube broadcast as follows:

1. Captions are pulled from the YouTube transcript track.
2. An LLM (Anthropic Claude) cleans the transcript — removes ums, restores paragraph breaks, fixes obvious caption errors — and structures it into reading sections. It does not add claims that weren't in the video.
3. Named entities (people, organisations, places, events, documents) are extracted with their context. Each entity gets its own archive page on this site.
4. The TL;DR, FAQ, and primary-topic strings are generated from transcript content only, with explicit instructions to preserve host framing on speculative material.
5. The page is written to storage with the original YouTube title and an editorially-cleaner descriptive H1.

The broadcast video is the canonical source. If the page text and the video disagree, the video wins.

06 // Submission handling

Material submitted via the submission portal is reviewed but is not guaranteed to air. Source identity is held confidentially. If the broadcast does reference submitted material, the submitter is contacted at the address provided unless they explicitly request silence.