Capturing The Tracking Websites
Three jobs no API can do.
Sites with no public API at all — the map is the only interface, so a browser is the only client. Proving what a page showed on a date, as a picture rather than a paraphrase. And the one this investigation needs most: documenting, carefully, where data that was once public is no longer retrievable.
That last one has to be done properly or not at all. The claim "they scrubbed it" is the easiest thing in the world to say and the easiest thing in the world to disprove, and one bad example discredits every good one.
What we already know is different
Two changes are confirmed from the open-source pass, both checked 24 August 2026:
- ADS-B Exchange historical pages return HTTP 403 to the public. Several of the original screenshots behind this whole claim were taken from that site while it was open. The data did not disappear — it moved behind a subscription when the site commercialised. Anyone re-checking an old screenshot there today cannot, and may wrongly conclude something was removed.
- airplanes.live now returns HTTP 403 and asks you to email a description of your project first. It was open before.
Neither is evidence of anything about this case. Both are exactly the kind of thing that gets mistaken for it.
And one that cuts directly against a scrubbing reading: the free historical archive returns 403 for every date from 12 October to at least 15 December 2025, for every aircraft we tried — including one with no connection to this case at all. That band happens to cover the dates a tracker claims SU-BTT was photographed at Sharm el-Sheikh. It is an archive-wide condition, not a suppression, and reporting it as one would be false.
What APIs they have — read as: which sites can be captured at all
apis/browser_capture/code/targets.json holds 128 targets — one per (site, aircraft) across
8 sites and 16 aircraft, each with the URL, the ICAO hex, which investigation thread the aircraft
belongs to, and what that site is good for.
| Site | What it is worth capturing for |
|---|---|
| Flightradar24 | The aircraft-history table the original claims were screenshotted from. Free tier shows a limited window — record how deep that window goes, it is part of the finding. |
| ADS-B Exchange | Free map only; historical replay is behind the paywall. Capture the paywall itself. |
| FlightAware | Aggressive bot protection. If it blocks, record the block and stop. |
| RadarBox | An independent receiver network — a genuine cross-check, not another view of the same feed. |
| Planespotters / JetPhotos | Photographs with dates and places. A photograph of an airframe at a field on a date is a different kind of evidence from a transponder reading, and it is not vulnerable to the same failure modes. |
| adsb.lol globe | The free map over the same archive the open-source scripts pull as JSON. |
| airplanes.live globe | Its API is gated now; record whether the map still is not. |
How you use them
Driven by Claude Code's browser tools, following apis/browser_capture/p_get_data.mdx. There is no
scraper here on purpose — the point is to capture what a human visitor sees, rendered, with its
JavaScript run.
Every capture needs four things or it does not count: the screenshot, the page text, the URL and the UTC time of capture, and the HTTP status.
Alongside the screenshot, monolith freezes a whole page into one self-contained file that still
renders years later after a redesign:
monolith -o captures/<site>/<tail>/<utc>.html "<url>"
A screenshot proves what was displayed. Monolith preserves what was underneath. Use both.
Rules that are not negotiable
- Never log in, never defeat a paywall, never click through a terms gate, never create an account. A page that requires it is a finding, not an obstacle.
- Never trigger a JavaScript alert, confirm or prompt. A modal blocks every further browser command and the session has to be rescued by hand.
- Images go to the repo and are embedded by local repo path, never by an IPFS gateway URL. Images are tracked in git — an untracked image renders locally and 404s for every real visitor.
- Check the ban list before embedding anything.
The comparison is the finding
A capture alone says "this is what the page shows today". Paired with the Internet Archive snapshot
nearest the claim date — wayback.js in the open-source pass, which reads the CDX API and reports
snapshot counts, dates, HTTP statuses and content digests — it can say "this is what the page showed
then, and this is what it shows now."
Two things must never be blurred:
- NEVER SHOWED IT — no evidence the page ever carried this. That is not scrubbing. It is the absence of a claim ever having been supported.
- SHOWED IT AND NO LONGER DOES — a documented change in the public record, with dates on both ends.
And it still cannot tell you why
A redesign, a URL change, a retention window rolling over, a commercial paywall and a deliberate removal all look identical from outside. Nothing in this pass distinguishes them.
So every page built from these captures states which of those it can and cannot rule out. A page that alleges scrubbing without ruling out a redesign will be taken apart by the first person who checks — and it deserves to be.
What we got so far
No browser captures yet. apis/browser_capture/captures/ is empty and this section will say so
until it is not.
But the Internet Archive half of the job has been run, and it produced the most uncomfortable finding on this page.
Nobody archived the evidence
30 target URLs checked against the Internet Archive CDX API on 24 August 2026 — the tracking-site pages for the five Egyptian tails and T7-ELL across Flightradar24, ADS-B Exchange, FlightAware, RadarBox and Planespotters.
22 of the 30 have zero snapshots. Including both of these:
| Page | Snapshots | Range |
|---|---|---|
| Flightradar24 — SU-BTT | 0 | never captured |
| ADS-B Exchange — SU-BTT | 0 | never captured |
| Flightradar24 — SU-BND | 2 | Oct 2025 – Nov 2025 |
| Flightradar24 — SU-BTU | 1 | Aug 2022 |
| Flightradar24 — SU-BGM | 1 | May 2020 |
| Flightradar24 — SU-BTV, T7-ELL | 0 | never captured |
| ADS-B Exchange — SU-BTU / SU-BTV / T7-ELL | 3 each | 2022 – 2023 |
| ADS-B Exchange — SU-BND / SU-BGM | 2 / 1 | 2022 |
| FlightAware — all five | 0 | never captured |
| RadarBox — all five | 0 | never captured |
| Planespotters — all five | 0 | never captured |
SU-BTT is the lead aircraft of the entire following claim, and neither its Flightradar24 aircraft-history page nor its ADS-B Exchange page was ever archived — while all four of the other Egyptian tails have at least one snapshot of one or the other. Not before the claim, not during it, not after.
The two SU-BND snapshots that do exist are both dated after the October 2025 broadcast that launched the claim. They record the reaction, not the evidence.
Every ADS-B Exchange snapshot predates the claim by two to three years, so none of them shows what a tracker was looking at in autumn 2025 either.
What this means, stated carefully. It does not mean anything was removed. It means that for the aircraft at the centre of this case there is currently no independent record of what the tracking sites displayed at the time the claims were made — so "the page changed" cannot be tested, in either direction, for the tail that matters most.
And one caveat that must travel with the table. The CDX API matches exact URL strings. A page archived under a slightly different URL form would not appear here. "Zero snapshots for this URL" is not the same as "never archived anywhere", and this table should be re-run against URL variants before anybody leans on it.
Raw responses: apis/public_open_source/data/wayback/.
What we did NOT get
- No captures at all, of any site, on any aircraft. 128 targets, none done.
- No URL-variant sweep. The zero-snapshot result above needs re-running against trailing slashes, alternate paths and the sites' older URL schemes before it is quotable.
- No copy of the original screenshots the claims were read off, with their own provenance — which post, which date, which account. Given that the Archive holds nothing for SU-BTT, those screenshots may now be the only record of what the page showed, which makes collecting them with their provenance the single most urgent task on this page.
- No record of which tracked tails, if any, appear on a vendor blocking list.
- No capture of the sites that are now closed — the ADS-B Exchange historical pages and the airplanes.live API — as they appear today, showing the paywall and the gate. Those are quick and should be done first.