Her story, over time
D1Who did the searching: a young escort -> a hotel/guest list -> an automated AI.
strong
First, 2026-07-09
“I might say it because I'm literally driving to pick this person up at the airport and I just need to check really quick what they look like ... For James Adams, they're looking September 4th through September 6th.”2026-07-09, 34:25
Jul 9: the searcher is an escort checking a face before an airport pickup.
Later, 2026-07-13
“It's much more likely that a military person, most likely an escort, was conducting that search so that they'd be able to visually identify the person that they were charged with going to pick up.”2026-07-13, 1:56:31 on Rumble
Jul 13 (Stew show, Ana answering Stew): a military escort.
More of her statements on this
Also, 2026-07-13
“they grab their phone and usually on the way to the airport, they are Googling that person”2026-07-13, 1:02:14 on Rumble
Same show: a person on a phone.
Also, 2026-09-19
“this list of individuals who would appear to have potentially been at the Candlewood Inn and Suites on Fort Huachuca”2026-09-19, 1:39:14
Sep 19: the list is now a hotel/guest list.
Also, 2026-10-03
“If they didn't know where the Google trend searches were coming from, how could it be coming from a person?”2026-10-03, 12:23
Oct 3: not a person at all.
Also, 2026-10-03
“what I realized is it's an automated system ... the new API leaves the record on Google.”2026-10-03, 26:29
Oct 3: an automated API system.
Also, 2026-10-03
“This isn't some data scraper ... This isn't some super curious gate guard. This isn't even, I believe, an escort ... for the longest time, my best guess has been there's such a super anal secretary who's like,”2026-10-03, 1:28:05
Oct 3, same stream: a fourth explanation (a secretary).
same blips, three different explanations (escort on a phone, hotel list, AI on a schedule) and now a fourth ("secretary") in the same stream that announces the AI. The data never changed; only the story did.
Honest limit: Verified: transcripts read at bmjVy 34:25, STEW 1:02:14 + 1:55:01-1:57:00, -rIa6 1:39:14, wmLl 12:23, 26:29, 1:28:05. [STEW-W] Whisper does not split speakers - speaker named by context, confirm by ear before airing.
D10Who wrote "ISRTF-PLTR": "that is how I determined it" -> "why would the AI think...?"
strong
First, 2026-09-14
“are we going to call him the, ISRTF-PLTR Alexander C. Karp, CEO. No, that's what a badge would read. I mean, that is how I determined what it was going to be that he would be being looked up as. And trust me, we put in every variation that we could come up with.”2026-09-14, 1:15:08
Sep 14: she built the label herself.
Later, 2026-09-26
“I want you to also note the ways that I'm listing folks ... no one should think that Johnny Moore is a part of a Palantir recon task force. And yet the way that he appears, somebody thinks that he's related to them.”2026-09-26, 27:17
Sep 26: 'somebody thinks'.
More of her statements on this
Also, 2026-10-03
“why is it that they're using like this exact phrase? Why is it that they're using this exact terminology?”2026-10-03, 24:56
Oct 3: the label is presented as the system's.
Also, 2026-10-03
“why would Gen AI's system think that Erica Kirk”2026-10-03, 58:17
Oct 3.
Also, 2026-10-03
“why would this API log think that Andrew Colfatt had anything to do with Palantir?”2026-10-03, 1:00:11
Oct 3.
on Sep 14 she says she built the label herself and tried every variation; on Oct 3 she presents the same label as something a government system chose.
Honest limit: Verified: an6d 1:14:02-1:15:30; nxbu 0:27:17; wmLl 24:56, 58:17, 1:00:11.
Correction: The Sep 14 "that is how I determined it" clip stays the core receipt. Don't say the ISRTF/Palantir link to these people first appeared that week.
D2Same list: "searches" on Sep 26 -> "these are not searches" / "API digital visitor log" on Oct 3.
strong
First, 2026-09-26
“these dates are not random. These searches are not random.”2026-09-26, 44:00
Sep 26: the rows are searches (HER claim that they are 'not random').
Later, 2026-10-03
“These are not searches, you know, no one, it's not like one person is here on, you know, July 2nd and then the other is here. They are three specific dates.”2026-10-03, 37:55
Oct 3: the same rows are 'not searches'.
More of her statements on this
Also, 2026-10-03
“all showing up on the same day through the API digital visitor log”2026-10-03, 1:09:05
Oct 3: now an 'API digital visitor log'.
the rows are Google Trends searches she ran; a week later the same rows are "not searches" but an API log.
Honest limit: Receipt: same Canva "FORT HUACHUCA LIST" rows both streams. "API" appears in ONE of her uploads (Oct 3, 21x); "GenAI/Gemini/grounding" about this topic only Oct 3 (our count across 10 Whisper transcripts + 69 caption tracks, 2026-10-06). Verified: nxbu 43:51-44:30, wmLl 37:55 (Whisper AND captions agree), 1:09:05. Any "not random" quote is labelled HER claim.
D3"They wiped it" vs "it's a sampling method" - one day apart, then back again.
strong
First, 2026-07-16
“let's do it because they have wiped some of them”2026-07-16, 37:25
Jul 16 collab: 'they wiped'.
Later, 2026-07-16
“hypothetically, yes, there are times when the data is captured in a small slice of the pie because it is done through a sampling method ... or if it's more likely that it's just, you know, this randomization has caused this.”2026-07-16, 1:05:39
Same stream, after the live re-run failed: she names sampling / randomization.
More of her statements on this
Also, 2026-07-17
“they started deleting the results like within 24 hours of me going on um on Stew ... they wiped so many results that they actually they messed up”2026-07-17, 44:02
Next day, Jul 17: back to 'they wiped'.
Also, 2026-08-20
“as soon as I went on st[ew] within 24 hours they wiped those searches ... you don't rush to cover something up unless you're kind of on to something.”2026-08-20, 22:06
Aug 20: still 'wiped'.
Also, 2026-09-14
“as we know, they love to go in and manipulate the data on the backend.”2026-09-14, 1:14:02
Sep 14: 'manipulate the data'.
Also, 2026-07-16
“Does not delete historical search data from Google Trends based on age, but it does strictly filter out ...”2026-07-16, 2:13:05
Counter-receipt on her own screen: she reads Google's AI answer.
Also, 2026-10-03
“the new API leaves the record on Google”2026-10-03, 26:29
Oct 3: if they wipe within 24 h, how is a Sept 2025 "record" still there in Oct 2026?
she named the real cause (sampling) on air, then went back to "deleted by the government" the next day and kept it.
Honest limit: Verified: a1Obe 0:37:08, 1:03:41-1:06:00, 2:12:32-2:13:30; OEsx 0:43:41-0:45:38; 8Obr 0:21:43-0:22:37; an6d 1:14:02; wmLl 26:29. a1ObeHOaq1k is OUR collab (@danksterintel is on it).
D4Erika Kirk: "wiped through Sept 13" -> found on Sept 6 -> July -> "Saturday" vs Snow's Sept 8.
strong
First, 2026-07-17
“they had deleted so much that they got rid of any searches for Erica Kirk from Arizona ... I think that they had wiped every single search result for Erica through September 13th.”2026-07-17, 44:30
Jul 17: every Arizona search for Erika wiped through Sept 13.
Later, 2026-09-19
“I have been trying to find this woman ... Is she ever going to appear? Could Mitch have just misremembered?”2026-09-19, 7:09
Sep 19: still not found.
More of her statements on this
Also, 2026-09-19
“oh my gosh, I finally found her ... the final last piece of puzzle, which is what I found last night ... it's Erica with a C”2026-09-19, 1:59:28
Sep 19: found (with a C).
Also, 2026-09-26
“JDA Erica Fransve appears on that list ... Same day, that is the last day that Charlie is speaking at SAS”2026-09-26, 27:17
Sep 26: now also a July hit (Student Action Summit, Tampa, July 2025). Her own sheet row: "EVENT: JDA ERIKA F ON FORT H LOG" between the Tampa SAS rows [FRAME nxbuNswDPpA/001660.jpg].
Also, 2026-10-03
“she showed up on the Saturday list as part of our weekenders”2026-10-03, 58:17
Oct 3: Saturday = Sept 6 2025.
Also, 2026-10-03
“saw individuals start to come to the base, including Erica, on September the 8th ... although the meeting took place on the 9th, the expected digital log would appear on the 8th.”2026-10-03, 1:01:56
Oct 3: Snow saw her on Sept 8; by her own rule the log should show the 8th.
by her own rule the "log" should show Erika on the 8th; it shows the 6th (and, on Sep 26, July). Two months earlier she said every Arizona search for Erika had been wiped through Sept 13.
Honest limit: Verified: OEsx 0:44:00-0:45:18; -rIa6 0:07:09, 1:59:28; nxbu 0:25:46-0:27:40 + frame; wmLl 58:17, 1:01:56.
D5DIA director Adams: "star of the show" -> "outgoing ... normal official business".
strong
First, 2026-07-13
“all of those individuals, including the star of the show, who would be James Adams, was being Google searched by his image in the immediate days prior, right up until, I believe, the day before the meeting.”2026-07-13, 1:53:51 on Rumble
Jul 13 (Stew show, Ana): Adams is the star of the show.
Later, 2026-07-17
“Lieutenant General Adams, that's another person who is identified in the Google Trends as having been there. He's brand new to his role.”2026-07-17, 28:31
Jul 17: 'brand new to his role'.
More of her statements on this
Also, 2026-09-26
“the date is going to be August 10th. It's the positive search result.”2026-09-26, 18:37
Sep 26: sheet row "DOD DIA DIR (OUTGOING) | ADAMS, LT GEN JAMES H | FH | Aug 10, 2025" and again "Sep 5, 2025" [FRAME nxbuNswDPpA/001091.jpg, t=1092].
Also, 2026-10-03
“the DIA's outgoing director ... It is, in fairness, very possible that ... Those two guys were just there for normal official business ... that doesn't qualify for like a big, big deal to me.”2026-10-03, 56:14
Oct 3: slide "SEPTEMBER 5, 2025 James H Adams DIA OUTGOING DIRECTOR" [FRAME wmLljWb4bcM/003379.jpg]; now 'normal official business'.
"brand new" in July, "outgoing" in September; the headliner of the July meeting theory is "normal business" in October.
Honest limit: Verified: STEW 1:53:23-1:54:32; OEsx 0:28:05-0:28:49; nxbu 0:18:09 + frame; wmLl 56:14 + frame.
D6Justin Davis ("the inside man"): Sept 7 -> Sept 6 -> Sept 8, on a common plain name.
strong
First, 2026-07-09
“in the days immediately prior ... Maybe. Allegedly ... All the way up to September 7th.”2026-07-09, 1:20:49
Jul 9 chart [FRAME bmjVyIJelZc/004837.jpg]: plain term "Justin Davis", Tucson (Sierra Vista), 8/1/25-9/10/25, Image Search - spike measured at Sep 7 (axis ruler 37 px/day from Aug 16 / Aug 31 labels).
Later, 2026-07-13
“I was surprised to see that on, let's see here, September 6th, his image was being specifically searched”2026-07-13, 1:39:11 on Rumble
Jul 13 (Stew show, Ana): Sept 6.
More of her statements on this
Also, 2026-10-03
“The inside man. That's Justin Davis ... on that same log and on that same day, there is an inside man.”2026-10-03, 1:28:05
Oct 3 slide "SEPTEMBER 8, 2025 Justin Davis INTEGRITY SOLUTIONS (THE SECURITY GUY) AND THE DRIVER OF THE SUV" [FRAME wmLljWb4bcM/005356.jpg]: Sept 8.
Also, 2026-07-09
“It's a relatively common name, and so I can't really make any inference off of the fact that someone ... in Sierra Vista ... they were looking up Bradley Hansel.”2026-07-09, 26:33
Her own rule, Jul 9 (on "Bradley Hansel"): a common name can't be used.
three different days for the most serious accusation she makes, built on a plain common name - the exact kind of hit she said can't be used. (Stew separately says "September 8th and 9th", rumble 2:10:56 - Stew's words.)
Honest limit: Verified: frame measured; bmjVy 1:20:49, 0:26:03-0:27:48; STEW 1:39:11; wmLl 1:28:05 + 4 frames.
D7Hansell "control": July + Aug 24 -> May 25 live -> "Sept 8, there he was".
medium-strong
First, 2026-07-09
“It looks like Bradley Hansel was being searched in his capacity as OUSDI during July. But I see this later search, August 24th ... That's a lot later than all the other searches.”2026-07-09, 31:40
Jul 9: latest Hansell hit is Aug 24.
Later, 2026-07-16
“Brad Hansel, O-U-S-D-I. Perfect. Okay, so that didn't show ... oh we hit on bradley just bradley full regular uh oh usdi ... image search and that was may 25th”2026-07-16, 1:38:55
Jul 16 live: 'didn't show', then May 25.
More of her statements on this
Also, 2026-10-03
“So this was my starting place for developing the entire meeting list. I knew that Brad had been there. He confirmed it himself. I also knew that if this method ... were accurate, then what I would expect to see is Brad Hansel there on September 8th. And there he was.”2026-10-03, 1:03:57
Oct 3: the 'control' is Sept 8.
Also, 2026-07-16
“he was at that meeting. Well, maybe allegedly. allegedly.”2026-07-16, 29:58
Jul 16 on whether he attended: 'allegedly' -> Oct 3 "He confirmed it himself" (no source shown).
in July her latest Hansell hit was Aug 24 and the live run gave May 25; the Sept 8 "control" she calls her starting point is not in any July material.
Honest limit: Verified: bmjVy 0:31:09-0:31:40; a1Obe 0:29:58, 1:38:55-1:39:30; wmLl 1:03:57.
Correction: Debate prep: concede the visit. A real visit still doesn't make a Trends blip a real search.
D8Brian Harpole: "searched in the days prior" -> "could also mean that he wasn't there".
medium
First, 2026-07-09
“I tried Brian Harpole Integrity. No. Brian Harpole Integrity Security Solutions. No. Brian Harpole Security. Yeah ... on August 26th, Starting August 19th”2026-07-09, 1:16:26
Jul 9: try-until-it-hits sequence, then a hit.
Later, 2026-07-09
“Brian Harpo's security was being searched in the days prior. John Ragland, Brian Harpoon. I think someone just misspelled it. I think it was an autocorrect.”2026-07-09, 1:28:56
Jul 9: "Brian Harpoon" counted as him.
More of her statements on this
Also, 2026-10-03
“Brian Harpo's name did come across, but not on that same day as the other individuals ... Perhaps he only required a brief and then stayed on post ... And it could also mean that he wasn't there.”2026-10-03, 2:13:53
Oct 3: the July hit becomes a non-hit.
a July "hit" (including "Brian Harpoon" counted as him) quietly becomes a non-hit when it doesn't fit the October three days.
Honest limit: Note: Harpole is a plaintiff in Harpole v. Owens - handle with care on air. Verified: bmjVy 1:16:26-1:17:40, 1:28:56; wmLl 2:13:15-2:15:10.
D9"Hard data" / "not proof" / "who's to say" / "high degree of certainty".
medium
First, 2026-07-09
“I am a very, very big fan of hard data because again, it changes the onus from being on the investigator to being on the individual ... that's up for them to tell us because now we already know.”2026-07-09, 12:32
Jul 9: 'hard data ... we already know'.
Later, 2026-07-13
“the hard data that we are able to generate based off of our trends, searches, specifically the image searches, is extremely edifying”2026-07-13, 18:55
Jul 13: 'hard data'.
More of her statements on this
Also, 2026-07-09
“It's not proof, but it's certainly, I do think it buoys the theory.”2026-07-09, 1:28:56
Jul 9: 'not proof'.
Also, 2026-09-19
“So I feel none of it is proof of anything, but it's suspicious.”2026-09-19, 2:11:05
Sep 19: 'none of it is proof'.
Also, 2026-10-03
“does that mean that every single person actually showed up? ... Who's to say?”2026-10-03, 24:24
Oct 3: 'who's to say'.
Also, 2026-10-03
“now that I would say that I have a very high degree of certainty that we know who was there”2026-10-03, 2:29:16
Oct 3: 'very high degree of certainty'.
Also, 2026-10-03
“we're not accusing anyone of anything”2026-10-03, 1:19:59
Oct 3: 'not accusing anyone'.
Also, 2026-10-03
“do I think that I want all of those people in prison? Yes, I would. I would like them indicted.”2026-10-03, 2:27:00
Oct 3: 'indicted'.
"not proof" when challenged, "hard data / we already know / indicted" when presenting. Same data throughout.
Honest limit: MEDIUM (pattern, not one line). Verified at each timestamp above. CAUTION: "you're caught. You did this" (wmLl 1:45:41) is about the Hades jet flying over Candace/[name excluded]/Stew's houses, not about the visitor list. Do not pair it with "not accusing anyone" should carry this context note).
Searches that failed, live or admitted
E1Jul 16 live re-run of her Stew chart - "the searches are gone".
strong
First, 2026-07-16
“this is our search here and you can see a very big difference ... we only have two ... Andrew ... July 6 ... Johnny ... August 2nd or 3rd”2026-07-16, 1:03:41
@danksterintel, running the re-run live.
Later, 2026-07-16
“now it's gone and now that exact time frame when they would or would not have been possibly there ... they're gone I mean the searches are gone”2026-07-16, 1:03:41
Ana: the searches are gone.
More of her statements on this
Also, 2026-07-16
“it is done through a sampling method”2026-07-16, 1:05:39
Then: sampling (see D3).
the one documented cross-day re-run with the same settings did not reproduce.
Honest limit: Verified: a1Obe 1:00:50-1:06:00. First quote is @danksterintel speaking, not Ana.
E2Jul 16 live: "Brad Hansel O-U-S-D-I ... that didn't show" -> different spelling -> May 25.
medium
First, 2026-07-16
“Brad Hansel, O-U-S-D-I. Perfect. Okay, so that didn't show ... oh we hit on bradley just bradley full regular uh oh usdi ... image search and that was may 25th”2026-07-16, 1:38:55
Live: no hit, then a different spelling hits May 25 (see D7).
Later, 2026-07-16
“i added johnny moore okay so no data”2026-07-16, 1:00:50
Earlier in the same stream: no data.
Honest limit: Verified: a1Obe 1:00:50, 1:38:55. @danksterintel is on this stream.
E3Jul 20 live on her own stream: the exact address misses; an unwanted hit "doesn't mean anything".
medium
First, 2026-07-20
“see how there's not a positive result for this. The exact address [street address redacted]”2026-07-20, 1:42:05
Live: the exact address gives no result.
Later, 2026-07-20
“I was not expecting to see a positive result right there. So, I'm a little bit not happy ... we have a March 2nd, a February 13th, and we have a April 24th. Now, it doesn't mean anything.”2026-07-20, 1:57:06
Live: an unwanted hit 'doesn't mean anything'.
when the method lands on someone she doesn't suspect, it "doesn't mean anything"; when it lands on a suspect, it's a visit. Same tool.
Honest limit: Verified: i9oCB 1:42:05, 1:56:46-1:58:09. The street address in the first quote is redacted here for privacy (source has it in full).
E4Oct 3: Cabot Phillips and Harpole not found -> "dead ringer" + "absence of evidence".
medium
First, 2026-10-03
“We did not see Cabot Phillips ... the person that we did get is honestly a dead ringer ... Martin Kindle ... absence of evidence is not evidence of absence”2026-10-03, 2:13:15
A miss becomes a lookalike and 'absence of evidence'.
Honest limit: Verified: wmLl 2:13:15-2:14:00.
E5CORRECTION to: the Sept 5 "I don't see him" is NOT a live search.
First, 2026-09-05
“of course I searched for him in the beginning ... I search for everybody ... I was like, I don't see him. I don't see him ... Well, I didn't realize he would be there as a guest of our DIA or DGI”2026-09-05, 2:33:04
Sep 5: her telling the story of looking for David Barnea (repeated in QKtLFt5c1Rw ?t=1185 caption).
Honest limit: It belongs in G (re-labelled until it hit), see G6. Don't call it a failed live search on air. Verified: AXTz 2:31:23-2:34:31; QKt 0:19:30-0:19:58.
Correction: CORRECTION to: the Sept 5 "I don't see him" is NOT a live search.
2026 titles on 2025 visits
F1Her own words: the titles and jobs come from her lookups "now".
strong
First, 2026-10-03
“This new computer system is forced to perform a live Google image search by title from the .gov directory. The same .gov directory where I was getting everybody's titles”2026-10-03, 16:34
Oct 3: she got the titles from the .gov directory.
Later, 2026-10-03
“All I had to do was just grab the individual's names out of the DOD directories ... the individual, how they were Googled, what their role is, what their title is, where they're working now ... and, of course, the most important and fun part, the search.”2026-10-03, 1:35:58
Oct 3: 'where they're working now'.
"where they're working now" = 2026. Any 2026-only title or employer on a 2025 row came from her, not a 2025 system.
Honest limit: Verified: wmLl 16:34, 1:35:58.
F2Ondas: her own slide says nobody worked there yet; her sheet and search term say ONDAS.
strong
First, 2026-09-14
On her screenNO ONE ON THIS LIST WAS EMPLOYED AT ONDAS DURING THIS TIMEFRAME. ONDAS WAS NOT YET PARTNERED WITH PALANTIR.2026-09-14, 42:29
Sep 14 slide [FRAME an6dnXjnYwg/002549.jpg]. Same chart, Trends term typed: "Yoav Har Even ONDAS" (window 1/1/25-9/9/25) and "D-DIA David Barnea".
Later, 2026-09-14
“DAVID BARNEA Fmr. Director of Mossad 2021 - June 2026: Director of Mossad”2026-09-14, 40:19
Sep 14 slide [FRAME an6dnXjnYwg/002419.jpg]: 'former' is a 2026 fact.
More of her statements on this
Also, 2026-10-03
On her screenONDAS*/MOSSAD - FMR DIR | BARNEA, DIRECTOR DAVID | FH | Aug 5, 20252026-10-03, 1:36:35
Oct 3 sheet [FRAME wmLljWb4bcM/005795.jpg]; also "ONDAS/MOSSAD - FMR DIR / ISRTF | BARNEA | Sep 8, 2025"; "ONDAS/PLTR - FMR CO-CEO ... LUGASSY, OSHRI | Sep 4, 2025"; "ONDAS/PLTR - FOUNDER / ISRTF-PLTR* | KARP | Sep 8, 2025"; "ONDAS/PLTR* / 8VC FOUNDER | LONSDALE, JOE | Sep 4, 2025".
by her own slide, "ONDAS" and "former Mossad director" are 2026 facts. They're typed into a 2025 search term and printed on 2025 "log" rows.
Honest limit: Verified: 3 frames viewed by eye.
F3Cliff Sims as "JD Vance's National Security Advisor" on Sept 8, 2025.
strong
First, 2026-09-05
“Mr. Cliff Sims. That is the new National Security Advisor to J.D. Vance. Cliff Sims, he is another person who is on our Fort Banana list.”2026-09-05, 1:51:41
Sep 5; slide text: "CLIFF D. SIMS * Nat Sec Advisor ... Newly named" [FRAME AXTzDnziLLk/006710.jpg, OCR].
Later, 2026-10-03
“Clifford D. Sims, as in J.D. Vance's national security advisor”2026-10-03, 1:12:13
Oct 3 slide "SEPTEMBER 8, 2025 Clifford Sims JD VANCE'S NATIONAL SECURITY ADVISOR (THE NATSEC BOSS)" [FRAME wmLljWb4bcM/004368.jpg, OCR].
11).
Honest limit: STRONG (with gap). GAP: we have her slide label, not the Trends search string she typed for Sims.
F4Adams "DIA director (outgoing)" in Aug/Sep 2025.
medium
First, 2026-07-09
“director of the Defense Intelligence Agency”2026-07-09, 33:53
July: 'director of the Defense Intelligence Agency' and 'brand new to his role'.
Later, 2026-10-03
On her screenSEPTEMBER 5, 2025 James H Adams DIA OUTGOING DIRECTOR2026-10-03, 56:19
Sep 26 sheet + Oct 3 slide.
Honest limit: MEDIUM - re-verify. Re-check the DIA roster (Kruse removed Aug 2025; who led DIA Sep 2025) against an official source before this goes on screen.
F5"D-DIA David Barnea" - a U.S. DIA prefix on Israel's Mossad director.
medium
First, 2026-09-14
“D-DIA David Barnea”2026-09-14, 42:29
Sep 14 Trends term [FRAME an6dnXjnYwg/002549.jpg].
Later, 2026-09-06
“David Barnea DGI”2026-09-06, 4:34
Sep 6/7 terms "David Barnea DGI", "D-DIA David Bar..." (trends_frames.json: p-zztTg-kWE 274, QKtLFt5c1Rw 786).
No year in which that was his title. Ties to E5/G6: she added it after "I don't see him".
Honest limit: Terms read from frames; the Sep 6/7 term is truncated on screen ("D-DIA David Bar...").
How the 43 and the three dates were picked
G1Hundreds searched, 43 kept, three dates picked after looking.
strong
First, 2026-10-03
“out of hundreds, hundreds and hundreds of individuals searched, a list of 43 people ... 38 of those individuals all showed up on the same three days. There's only five outliers ... So as I narrowed in on these three dates, August 28th, September 4th, and September 8th”2026-10-03, 29:19
Oct 3: the denominator is 'hundreds and hundreds'; the dates were narrowed in on.
Later, 2026-10-03
“every single individual, along with tons, tons, tons more people in the .gov directory, the only ones that show up on September the 8th are the bosses.”2026-10-03, 1:04:54
Oct 3.
Honest limit: Say "not on any list she has SHOWN".
G2"Not picking and choosing ... because so many people show up on so many different dates."
strong
First, 2026-10-03
“these dates are not random [HER claim]. They're not like all over the place, right? Like we're not like picking and choosing who to show because there's so many different people that show up on so many different dates now.”2026-10-03, 1:37:56
Oct 3: self-undermining - many people on many dates.
Honest limit: STRONG (self-undermining). "[HER claim]" is the research file's label, not her words.
G3"I always include" positives.
strong
First, 2026-07-13
“I started out all the way back as far as January. And then I narrowed the search from there. Anytime that there is a positive result, then I always include it.”2026-07-13, 1:10:42 on Rumble
Jul 13 (Stew show, Ana): positives always included.
(Misses aren't counted.)
G4Narrow the window until it "generates" a hit.
strong
First, 2026-07-20
“there was not enough interest overall for the past 2 years for it to actually generate a positive result. It is only once we start to narrow down.”2026-07-20, 1:23:12
Jul 20.
Later, 2026-07-16
“i start with the year and then i just kind of observe ... and then i go in”2026-07-16, 1:38:04
Jul 16 collab.
Honest limit: Collab = @danksterintel is on it.
G5Every variation, every misspelling, every person.
strong
First, 2026-09-14
“we put in every variation that we could come up with”2026-09-14, 1:15:08
Sep 14.
Later, 2026-09-19
“I sat down and went through every single DOD official, every single, every single person that I could humanly come up with”2026-09-19, 1:59:28
Sep 19.
More of her statements on this
Also, 2026-07-16
“i try all of the i even do common misspellings”2026-07-16, 1:39:22
Jul 16 collab.
Also, 2026-07-20
“once you start getting like pretty positive results, you have to check common misspellings”2026-07-20, 1:30:10
Jul 20.
Also, 2026-08-03
“even though they've managed to go back and wipe a bunch of it, I can find it under the way that the person misspelled it ... They're forgetting to wipe them.”2026-08-03, 41:09
Aug 3.
G6Re-label until he appears (Barnea).
strong
First, 2026-09-05
“I searched for him in the beginning ... I don't see him ... Why am I not seeing him? Well, I didn't realize he would be there as a guest of our DIA or DGI”2026-09-05, 2:33:04
Sep 5 -> term "D-DIA David Barnea".
Honest limit: See E5: this is NOT a live search.
G7Erika found by stepping through spellings.
medium
First, 2026-09-19
“first I had to realize ... maybe Erica was there. Then I had to realize Erica maybe was there as a guest to Palantir. Then ... the final last piece of puzzle, which is what I found last night ... it's Erica with a C, but it's still Erica Leigh and Frantz Faye”2026-09-19, 1:59:28
Sep 19.
G8Harpoon counted as Harpole.
medium
First, 2026-07-09
“Brian Harpo's security was being searched in the days prior. John Ragland, Brian Harpoon. I think someone just misspelled it. I think it was an autocorrect.”2026-07-09, 1:28:56
Jul 9.
Honest limit: See D8 and frame item FR3; Harpole is a private plaintiff.
G9Total tries: never stated.
No transcript gives a count of searches/terms/variants. Our searches for "how many searches", "thousands of names", etc. found nothing. Ask her on air (denominator).
Honest limit: .. I'm gonna need like six" (wmLljWb4bcM ?t=5811) is the closest thing to a count of her hits - never state 600 as fact.
Other
H1"It's like sampling ... if it's on there, the search happened."
strong
First, 2026-08-04
“it's like sampling ... But long story short, if it's on there, the search happened and that's like, you know, to me the most important indicator.”2026-08-04, 2:23:23
Aug 4, live with @danksterintel (@danksterintel supplies the word "sampling").
Honest limit: Note @danksterintel is on this clip.
H2"This isn't just random noise."
strong
First, 2026-10-03
“as you guys can see, this isn't just random noise. I really do feel comfortable saying that.”2026-10-03, 1:17:58
Oct 3 (vs Google). Already
Honest limit: Any "not random" line is labelled HER claim.
H3Common-name caution applied to Hansell, not to "Justin Davis".
strong
First, 2026-07-09
“It's a relatively common name, and so I can't really make any inference off of the fact that someone ... in Sierra Vista ... they were looking up Bradley Hansel.”2026-07-09, 26:33
See D6.
Honest limit: See D6.
H6Oct 3 "a heck of a control" vs Jul 9 "most likely".
medium
First, 2026-07-09
“We already have established that Brad Hansel is most likely someone who went to Fort Banana.”2026-07-09, 11:30
Jul 9: 'most likely'.
Later, 2026-10-03
“It's a heck of a control to know someone was going to be there.”2026-10-03, 25:27
Oct 3: 'a heck of a control'.
Honest limit: Not counted in H (they're contradicted by our tests, not by her): "only the bosses show up on Sept 8", Hansell re-pulls, Army-documented visits = 0-2.4).
H4The "new explorer is manipulated ... they still do wipes on the classic" - unfalsifiable.
weak
First, 2026-07-16
“the explorer seems to have a lot more like manipulation ... they'll still go in and do wipes and stuff like that on this um even on the classic explorer”2026-07-16, 59:51
Jul 16 collab.
Honest limit: WEAK; keep off screen unless asked.
H5Same move on flight data.
weak
First, 2026-09-08
“that website in particular, they do allow for data manipulation because ultimately they're owned by the usual suspects”2026-09-08, 4:09:09
Sep 8, on a viewer finding no flights.
Honest limit: WEAK, keep off screen. Pattern only; the "usual suspects" wording is coded - do not air.
What her on-screen charts show
FR1Her Stew Peters chart vs the live re-run one day later
strong
First, 2026-07-13
On her screenJohnnie Moore Jul 10/15; Kolvet ATK/T. Kolvet Aug 15/25; Kolvet Kairos ~Sep 7-8.2026-07-13, 0:00 on Rumble
Her chart on Stew (~Jul 15 2026) - dates as read in the frames file.
Later, 2026-07-16
On her screenMoore ~Jul 13 + ~Aug 2, Kolvet ATK ~Jul 6, Kairos nothing.2026-07-16, 1:04:40
Live re-run, same 5 terms + settings, Jul 16 collab (side-by-side at 1:04:40).
More of her statements on this
Also, 2026-07-16
On her screenit is done through a sampling method2026-07-16, 1:05:45
Ana ~1:05:45.
the only time anyone re-ran one of her charts on camera a day later, it did not come back.
Honest limit: Categories [a + e].
Correction: Use only dates measured off her chart frame, and label any spoken date as Stew's.
FR3"Brian Harpoon" - a misspelling spike counted as a search for Brian Harpole
strong
First, 2026-07-09
On her screenMeasured: Harpole Sec ~Aug 19, Harpoon ~Aug 26, Ragland ~Aug 31, Justin Davis ~Sep 7.2026-07-09, 1:19:47
Chart: "Brian Harpoon", "Brian Harpole Sec...", "John Ragland", "Justin Davis"; Tucson (Sierra Vista) AZ, 8/1/25-9/10/25, Image Search.
Later, 2026-07-09
On her screenThey looked up Brian Harpoon. They looked at I think an auto correct. They looked up Brian Harpole's security. So I don't think that there's two.2026-07-09, 1:18:50
Her words 1:18:50 (caption).
More of her statements on this
Also, 2026-07-09
On her screenThey were looking at Brian Harpoon. Maybe that's some children's cartoon ... Let me try other spellings ... I tried Brian Harpole Integrity. No. Brian Harpole Integrity Security Solutions. No. Brian Harpole Security.2026-07-09, 0:00
Cold open 0:00:00 (Whisper): her own try-until-it-hits sequence.
(1) a misspelled name blipped - exactly what a fake name does in our Daffy Duck tests; she counted it as him. (2) She shows her own try-until-it-hits sequence: Integrity (no) -> Integrity Security Solutions (no) -> Security (hit). (3) Harpoon and Harpole spiked a week apart, yet she calls them the same searcher.
Honest limit: Care: Brian Harpole is a private plaintiff (Harpole v. Owens). Show only her words + the chart. By her method he got "searched from" the area - never "was there". Category [g/h].
FR4District shopping - her own on-screen method for picking positives
strong
First, 2026-07-20
On her screenI just kind of go through them really quick ... center district. So, we had a July 11th. Now we have a July 7th. And I just take a note of that. Then I check in Haifa. Okay, nothing. Not really a huge deal. Then ... Jerusalem and nothing there. Then ... north ... We have a July 9th ... So we have a center and a north in our positive so far. And then we want to go to south.2026-07-20, 1:34:20
Her words 1:34:20-1:35:20 (caption); chart at 1:35:02 ("Victor Marx", North District, 5/1/25-9/9/25, Web Search -> tooltip Jul 9, 2025). Jul 20 2026, "FORT BANANA: The Digital Stakeout (Part 2)".
one name, six regions, different dates in each, misses are "not a huge deal", hits are "positives". That is the sieve, demonstrated by her, on camera. Every extra region is another lottery ticket.
Honest limit: Category [g].
FR5"I searched for Erica over and over ... never found" -> then her own "Palantir" label hits
strong
First, 2026-09-19
On her screenISRTF-PLTR Erica Lane Frantzve2026-09-19, 1:25:30
Slide 1:25:56 -> tooltip Sep 6, 2025 (Tucson (Sierra Vista) AZ, 1/1/25-9/9/25, Image Search).
Later, 2026-09-19
On her screenPlease say hello to our latest Google trend find. Erica Lane Fronfay Palunteer. Now, like I said, I have searched for Erica over and over and over and over. Couldn't find her ... I have been searching for this entire time and have never found...2026-09-19, 1:26:00
Her words 1:26:00 (caption).
months of misses, then a hit only on a string that starts with her own invented tag "ISRTF-PLTR". The tries that failed are not on screen; the one that hit is.
Honest limit: ) Categories [g + d].
FR2"DDI Bradley Hansell": she says August 24th, her own chart shows ~August 5
medium
First, 2026-07-09
On her screenBut I see this later search August 24th and I say that seems later. That's later than all the other searches. That's a lot later than all the other searches.2026-07-09, 31:40
Her words 0:31:40 (caption); chart on screen 0:31:18-0:31:50 (5 terms, Tucson (Sierra Vista) AZ, 5/1/25-9/9/25, Web Search).
Later, 2026-07-09
On her screenPurple "DDI Bradley Hans..." peak x=1456.5 -> Aug 5, 20252026-07-09, 31:18
Measured on the 1080p frame (source mp4 t=1878): axis May 1 x=345 / Jun 16 x=877 / Aug 1 x=1409 = 11.57 px/day. Control: the other four peaks read Jul 11, 12, 14, 18 - matching the known Hansell dates, so the ruler is right. No tooltip shown; no other chart on screen 0:31:18-0:32:18.
the date she read out isn't on her own chart. Small on its own; useful as "she reads dates by eye, live".
Honest limit: Honest limit: pixel-read, +/-1 day. She may have looked at a different load off camera. Category [c].
FR6The Las Vegas "data cluster": the 3 "concurrent" names are 0 on her own tooltip
medium
First, 2026-07-23
On her screenJul 23, 2025: Kolvet 100, Kent 99, Moore 0, Twiggs 0, Robinson 0.2026-07-23, 1:22:01
Chart: "Andrew Thomas Kolvet / JDA Kent Frantzve / Johnnie Moore / Lance Tyler Twiggs / Tyler James Robinson", Las Vegas NV, 5/1/25-9/9/25, Web Search. Tooltip values. Robinson peaks the day before, Twiggs/Moore the days after (each term its own day).
Later, 2026-08-04
On her screenTyler James Robinson was being concurrently Google searched on these same days ... And my feeling was I think we're past the point of random a little bit ... So my thought was I was like this doesn't feel random.2026-08-04, 1:00:00
Her words, Aug 4 2026 collab WITH DANKS (her channel, "AIR MEEMAW?? LIVE with Danks"), caption ~1:00:00. [HER claim - always label it as hers]
her own tooltip says the "cluster" names were zero on the day she points to; they are single blips on separate neighbouring days. Medium because "same frame of time" is loose wording.
Honest limit: Categories [c + h]. @danksterintel is on the Aug 4 stream.
FR7Reused screenshot, never re-pulled - the Vegas Jul 23 image across 4 uploads / 12 days
medium
First, 2026-07-23
On her screenIdentical image (same tooltip values, same cursor spot, same "A" avatar)2026-07-23, 3:14
b0H3ekbYPpk (Jul 23) @194, @4921, @5400, @5520.
Later, 2026-07-28
On her screenIdentical image2026-07-28, 1:04
hzh69oafqHA (Jul 28, edited "Intel Brief") @64.
More of her statements on this
Also, 2026-08-04
On her screenIdentical image2026-08-04, 1:00:29
1ur1d-1B7cA (Aug 4, with @danksterintel) @3629 + @3828.
Also, 2026-07-21
On her screen"JDA Erika/Kent/Andrew/Lori/Stacy" Jun 27 tooltip (Lori 77, Stacy 96, Las Vegas, 1/1/25-9/9/25)2026-07-21, 18:22
Also: pET5yLlWcB4 @1102 and @2548 (Jul 21) and b0H3ekbYPpk @5181 (Jul 23). The Barnea/DGI/Kruse chart is identical in p-zztTg-kWE @274 (Sep 6) and QKtLFt5c1Rw @786 (Sep 7) (edited minis, expected).
Also, 2026-07-23
On her screenlet me pull them up just to remind everybody2026-07-23, 3:14
Her framing: she says so herself (b0H 0:03:14).
So this is NOT "passing off old screenshots as new". The point is narrower: across 12 days and 3 shows she never re-ran it to see if it still came back. On the one occasion a chart was re-run live, it didn't.
Honest limit: Don't overclaim: re-showing a screenshot is normal. Never call it faking. Category [b].
FR8Same method, any state: a Hebrew-translated Utah address "searched" in Delaware
medium
First, 2026-08-20
On her screenwhy is it that on August 23rd, we are now seeing the search being performed in Delaware?2026-08-20, 54:50
Her words 0:54:50-0:55:40: matches it to the day an "Egyptian jet" flew to Delaware. Chart 0:55:16: Hebrew string (a Utah street address, not reproduced here) in Delaware, 1/1/25-9/9/25, Image Search -> tooltip Aug 23, 2025 = 100. Aug 20 2026, "TPUSA and A VERY Special Agent".
same recipe (translate, rare string, image search, small region), now a different state, matched to whatever date the story needs. Our control test did the same with Timpanogos hospital in Arizona (May 13).
Honest limit: Don't read the address on screen (she declined to "dox"; we keep it blurred-free but unspoken - show the chart only). Category [h]. The address itself is not in this file.