This is the record for September 28 to October 4, 2026, and it closes at about 18:42 UTC on Sunday 4 October. The 1,034-account roster is the frame we track, and what we collected inside it is smaller: 16,089 posts fall inside the window, 7,760 of them from X, 6,171 from Bluesky and 2,158 opening posts from the two boards, beside 490 news rows. Every finding below describes that monitored sample. The ledger holds 601 confirmed items, one to a url, and no script on it reached the top tier. In the same record, 57 items ask for violence or for removal by force. Of those, 26 are anonymous board threads, 2 reach us quoted in published reports and 29 come from named accounts, and most of the 29 rest on an inference recorded with the coding, usually the step that identifies the group behind an invader label, a pronoun or the word remigration. Some need no such step, among them a sitting party leader’s own demand. On 4 October Geert Wilders, who leads the Party for Freedom in the Netherlands, wrote that large parts of Belgium are already almost Arab and that above all criminals of migrant origin must be made to remigrate, closing “Wegwezen. Eruit!” (Get lost. Out!). The word he used for them, allochtonen, sorts people by origin and takes in Dutch citizens of migrant background, and we code the demand as a call for removal by force. All but five of the named calls sit in the script that describes native populations as replaced and offers the removal of migrants and their descendants as the answer, which is also this record’s peak at 3.6. Two accounts called that removal the moderate position and the moderate solution on 2 October, and a third asked that rioting French pupils be rounded up and removed with their whole families. Around those items the record ran on a small number of arenas. French high school protests over overcrowded classes, teacher shortages and run-down buildings escalated on 30 September, and on this instrument they became the largest script on the ledger: 102 items assigning the violence to people of migrant descent as a group, most of them posted by one account, visegrad24, which put one generational migrant label on scene after scene for four days. One account repeating its own caption is volume from a single carrier, and we read no movement into it. The polarization instrument recorded the same protests as its top card of the week, scored at 4.4 in its top watch tier and marked ignited: the French government, Nigel Farage and a cross-border set of right-wing accounts reading the violent turn of the blockades as France Unbowed, the hard-left party, inciting and using the pupils. On 30 September the co-pilot of Flydubai flight FZ1073 attacked the captain in the cockpit over Jordan, and the same day LauraLoomer asked that Muslim people be banned from piloting commercial planes; on 2 October the account asked that they be barred from elected office and federal work. Accounts in that arena worked to identify the co-pilot, whom the authorities have not named, and no confirmed item adds an identification. New York’s governor appointed a special prosecutor in the Cornell fraternity case on 1 October, and the case drew board threads calling men as a sex rapists beside a named post holding that a woman claims rape to save face “Every single time”. Rupert Lowe, a party leader in Britain, asked on 28 September whether the five men arrested near RAF Fairford had simply been handed a passport and labeled British, and on 4 October pledged that a Restore Britain government would reverse what he called the Islamification of the country. HOPE not hate reported a photo caption offering a final solution, posted by an activist who attended a Restore Britain branch meeting, and the code of conduct of a neo-Nazi chat licensing praise of Hitler; on 4 October a British account campaigning for the revival of the National Front answered the Mayor of London’s post on the 90th anniversary of Cable Street with “We march on”. Of the 34 ranked scripts, 13 stand in the second tier and 21 in the third, peak severity is 3.6 and the median is 2.8. The 12 scripts that only anonymous board threads carried have no impression count to read, so they are marked reach unmeasured and ranked in a list of their own, with 4 in the second tier, 1 in the third and 7 at the bottom. Every one of the 46 scripts carries the verdict no coordination identified, which means we found no bridge account, no named hub and no directive on any of them, and did not document independent origin either. Tiers rank scripts only, and the 26 items that match no script sit outside the ranking with no score. The 46 scripts carry 89,051,419 platform-reported impressions, about half of them on the French protest script.
The ledger holds 601 confirmed items, one to a url. The 46 scripts on it carry 89,051,419 platform-reported impressions, and the 26 items that match no script stand outside the ranking. The rungs run stigmatizing 324, hateful 117, biased 63, calling for violence 57, threatening 22 and glorifying violence 18, which accounts for all 601. The 57 calls are where this record’s weight sits, and the question to put to them is who made them. 26 sit on anonymous board threads, where the targets range across Black and non-white people, women, Indian people, white, European and American peoples, mixed-race people, political camps, Jewish people, Muslim people, Ukrainian people and people of migrant origin. Two reach the record through published reports: HOPE not hate reported on 29 September a photo caption by an activist who attended a Restore Britain branch meeting, “I have a final solution but it requires some concentration”, and Euronews reported on 3 October the standing demand of a movement called Save Europe Act for the compulsory or voluntary remigration of people it calls not integrated, a movement that a T-shirt sold as support for Ceuta helps fund. The other 29 come from ten named accounts, and most of them rest on an inference recorded with the coding, usually the step that identifies the group behind an invader label, a pronoun or the word remigration. TRobinsonNewEra supplies 11, nine of them through a stated inference. On 1 October it told Marine Le Pen to “Round them up and remove them all when you win”; on 3 October it wrote of the rioting pupils “Round them up, their whole families, and remove.”; and on 1 and 3 October it closed two posts on the French unrest with a demand to bring back the guillotine for people it calls invaders and the offspring of invaders. Steve_Laws_ supplies five, among them “Remigration or death. Those are the options.” on 2 October, and TheGloriousLion supplies five, among them a post of 4 October making support for mass remigration the test of a good Pope or a good person. On 2 October JaydaBF and Geiger_Capital, about five hours apart and with no quote, reply or shared hub between them in the record, called remigration the moderate position and the moderate solution. EvaVlaar wrote on 1 October “There is no integrating these kids. Remigration is the only solution”. ConceptualJames posted “Globalize the Nakba” on 2 October and wrote on 4 October that the protesting pupils “should mostly be deported and/or imprisoned, at a minimum.” AuronMacintyre answered a post saying that French-born rioters cannot be deported with “You can do every bit of this, and you must”, and DavidVance endorsed a post saying that the only peaceful solution is to send people back to their countries of origin. The 29th is Geert Wilders’s own. All but five of the named calls sit in the replacement and removal script. The five are the two guillotine posts, TRobinsonNewEra’s call to begin remigration from New York City and Dearborn, the Nakba post, and Steve_Laws_’s quoted line calling Jewish people a microbe whose elimination is the only remedy.
No script reached the top tier. Of the 34 ranked scripts, 13 stand in the second tier and 21 in the third; peak severity is 3.6, on the replacement and removal script, and the median is 2.8. Our anchors set harm potential at its top value where a script holds a concrete, explicit call for violence against a defined group, and a script there takes a second-tier minimum whatever its computed score. Ten scripts reach that value this week. Four reach the second tier on their computed scores, so the minimum changes nothing for them: the replacement and removal script; the script denying Muslim people a place, on TRobinsonNewEra’s call to begin removing Muslim residents from New York City and Dearborn and a post whose author says he holds the addresses of the people he wants raided; one account’s three posts closing on the guillotine; and the script describing Palestinian people as murderous by intent, on the Nakba post. Six take the tier from the minimum. Two of the six carry named accounts beside the boards: the script wishing for or demanding an end to the Jewish presence, which runs from board threads asking when it will be legal to hunt Jewish people in the streets to Steve_Laws_’s quoted microbe line and the caption HOPE not hate reported; and the endorsement of killing and beating members of political camps, six board threads and one RadioGenoa post praising a group it said was hunting anti-fascists in Madrid. The other four run on the boards alone and ask for the extermination of Black and non-white people, for killing or forced childbearing imposed on women as a class, for the annihilation of white and European peoples, and for harm and genocide against Indian people. The 12 board-only scripts have no impression count to read; they are scored on four measurable factors, marked reach unmeasured and ranked in a list of their own, with 4 in the second tier on the harm anchor, 1 in the third and 7 at the bottom. One script sits below the top harm value with a call inside it. The script reading Muslim presence as conquest holds a board thread at the calling-for-violence rung, a lament that no one has the resolve to kill Muslim people, and a call that can only be read through an inference does not set the top value, so that script’s harm potential sits at 4 and the thread keeps its rung.
The largest script by volume is the recasting of the French high school protests. The protests, over overcrowded classes, teacher shortages and run-down buildings, escalated on 30 September, when police detained 625 people at blockades affecting 356 schools. On this instrument they became 102 items assigning the violence to people of migrant descent as a group, and most of them are one account’s: visegrad24 put one generational label, “second and third-generation migrants”, on scene after scene from 29 September to 2 October, and on 4 October carried the same construction to a disturbance in Vicenza, Italy. RadioGenoa’s captions calling the young people African and Arab, or barbarians, and TRobinsonNewEra’s invader label supply most of the rest, and four board threads recast the unrest, three of them under a race label. The script carries about half of the impressions on the ledger, most of them on visegrad24’s posts. One account repeating its own caption is volume from a single carrier, and we read no movement and no coordination into it. The polarization instrument recorded the same protests as its top card this week, at 4.4 CPS in the W1 tier and marked ignited: the French government, Nigel Farage and a cross-border set of right-wing accounts read the violent turn of the blockades as France Unbowed, the hard-left party, inciting and using the pupils for partisan ends. The French protests carry two scripts on two instruments, a descent label on this one and a party blamed for the violence on the other, and some of the same accounts carry both. The Flydubai cockpit attack of 30 September anchors the script assigning violence to Muslim people as a group, where LauraLoomer carries nine posts, among them “Muslims should be immediately banned from piloting commercial planes”, and where TRobinsonNewEra asked the same day “Muslim pilots no problem at all ?”. LauraLoomer moved on 2 October to a ban on Muslim people in elected office and federal work, and on 4 October, after the Emirati prosecutor general called the attack terrorism, to an untrustworthiness trait of Muslim people drawn from Emirati and Qatari officials. Accounts in the arena worked to identify the co-pilot, whom the authorities have not named; no confirmed item adds an identification, and we do not name him. The Cornell case, in which New York’s governor appointed a special prosecutor on 1 October, anchors three board threads calling men as a sex rapists, and the same arena supplied named posts to two scripts about women, among them MyronGainesX on 30 September that a “woman claims rape to save face. Every single time”.
Serving officeholders, party leaders and one official body carry 13 items across 8 scripts in their own voices, and those 8 are the scripts whose mainstreaming factor sits at 5, which our anchors set at adoption by a politician or an official body speaking for itself. Geert Wilders supplies three. On 28 September, answering the Dutch prime minister’s remark that people attacking the rights of gay people, women and minorities would be hit in the face, he attributed half of the country’s sexual crimes to people of non-Western origin and wrote “ik wil ze alleen het land uit” (I only want them out of the country), coded threatening; on 1 October he wrote that African arrivals stab each other and the Dutch to pieces and are still let in; and on 4 October came the remigration demand. Rupert Lowe supplies three, among them his question whether the Fairford suspects had simply been handed a passport and his pledge to reverse what he called the Islamification of the country. Nigel Farage wrote on 1 October “No more immigrants from Afghanistan or Palestine when I’m in charge.” beside a promise to put women’s safety first, and on 2 October “We are inviting terror and horror into our communities.” Lee Anderson, a sitting member of parliament, called small-boat arrivals “boat loads of wrong uns” on 28 September. Maxime Bernier, a party leader in Canada, posted one statement in two languages on the National Day for Truth and Reconciliation that calls the residential-school genocide claim a hoax and says reconciliation requires an end to “the victim mentality, the fake white guilt, and the grifting.”, coded at the biased rung. The official account of the Senate Republican Conference listed “flood the country with illegal aliens” among what Democrats would do in power. The heaviest named carriers are activist and aggregator accounts: RadioGenoa appears in twelve scripts and TRobinsonNewEra in eleven, and together with visegrad24 they account for more than half of the named items on the ledger. By arena, 186 items are directed across borders, 171 at France, 109 at the United States, 77 at Britain, 12 at Italy, 10 at Germany, 9 each at Canada and Ukraine, 5 at Spain, 4 at Austria, 3 each at Greece and the Netherlands, 2 at Finland and 1 at Lithuania, which accounts for all 601.
The events that drew nothing we could code are recorded beside the ones that did. Jordan Bardella’s denial on 28 September of a report that private messages he wrote years earlier held antisemitic statements drew no confirmed item, and neither did Pope Leo XIV’s remarks the same day setting the far right apart from Christianity. The bail refusal of 29 September for the British activist charged over the slashing of a migrant dinghy in the Channel drew none, nor did the Clancy hearing the same day, Mark Carney’s remark about resuming a controlled increase in Canada’s population, or the immigration ruling of 30 September that the president of the Islamic Society of Milwaukee can be deported. The antisemitic flyers found at the Belleville synagogue on 1 October carry the hidden-control script in print, and no confirmed item mentions them. The terrorism charge of 2 October against the man already charged with the murder of Ann Widdecombe, with Nigel Farage among the alleged targets, drew no confirmed item before the close. The Vice President’s line at a Lakeland rally on 1 October, that a Florida Senate nominee who calls herself a socialist does not belong in America, did not clear our harm bar, which holds stock partisan idiom below it; the polarization instrument carries it as a card of its own.
Limitations this week
This record closes at about 18:42 UTC on Sunday 4 October, so the last five hours of Sunday UTC are outside it and Sunday’s items are counted to that hour. All 152 X handles on our collection list were read, and 139 of them carried in-window posts; the other 13 posted nothing inside the window that we captured, and we read them as quiet. The nightly X pulls of 1 and 2 October collected almost nothing, so the week’s X record rests on a walk of every listed handle and a backfill run on 3 and 4 October. A post deleted before that walk is absent, and the counters on early-week posts were read days after posting. Three handles stopped at the 200-post page cap, so their counts are floors on captured posts; none of the three carries a confirmed item. The two boards keep no archive we can return to, so the board items on this ledger are a lower bound on the qualifying threads those boards carried. No item on the ledger comes from Bluesky. The news layer holds 490 in-window rows and no general American politics beat, so the American arenas rest on the posts table with almost no coverage in our news table. We name no private individual on this page, including the Flydubai co-pilot, whom accounts in that arena worked to identify, and two men in a doorbell clip whom one account asked its followers to make famous.
Scripts by severity
Click a bar for the full scorecards.
The week's weather, day by day
One dot per real event. Filled dots mark the three events a scored script cites as its trigger; hollow dots mark the rest, and eight of those drew confirmed items inside scripts anchored elsewhere. This record closes at about 18:42 UTC on Sunday 4 October. Click for the full board.
The 601 items, by harm level
Click for the full ledger.
On the radar
Click a date for the full calendar and the patterns building toward it.
01 · The scripts
Five factors are scored from 1 to 5: reach, toxicity, coordination, mainstreaming risk, and harm potential. Each item that joins a script is assigned to exactly one, with overlaps adjudicated on the dominant predicate; the 26 items that match no script stand alone and appear on the ledger as individual hits. A script’s canonical claim is its shared core predicate, and escalation clauses, a removal remedy or an endorsement of force, are stated with the count of posts that carry them, which each script’s prose does. Their composite Ember Priority Score is shown on a chart running from 0 to 5; the axis starts at 0 for display, and the score itself cannot fall below 1. Click one for its scorecard. Peak severity this week is 3.6, the script describing native populations as replaced with the removal of migrants and their descendants offered as the answer, and no script reached the top tier; the median of the 34 ranked scripts sits at 2.8. Of those 34, 13 stand in the second tier and 21 in the third. The 12 scripts whose reach is unmeasured rank in a list of their own, 4 of them in the second tier, 1 in the third and 7 at the bottom. Our anchors set harm potential at 5 where a script holds a concrete, explicit call for violence against a defined group, and any script at 5 takes a second-tier minimum whatever its computed score. Ten scripts reach it this week. Four of the ten already compute into the second tier: the replacement and removal script, on posts asking that French pupils be removed with their whole families, on a post setting the options as remigration or death and on a party leader’s demand that criminals of migrant origin, citizens among them, be made to remigrate; the script denying Muslim people a place, on a call to begin removing Muslim residents from two named American cities; the guillotine demanded for French young people of migrant descent; and the script describing Palestinian people as murderous by intent, on a call to globalize the Nakba. Six take the tier from the minimum: the end of the Jewish presence wished for or demanded, the endorsement of killing members of political camps, and four scripts carried only by board threads that ask for the extermination of Black and non-white people, force against women as a class, the annihilation of white and European peoples, and harm and genocide against Indian people. Toxicity anchors at 4 for sustained othering and at 5 for explicit dehumanizing metaphor or called-for harm across the script, and 17 scripts sit at 5, among them the replacement and removal script, the conquest script, the vermin and insect labels for migrants, the parasite label applied to the Home Secretary, the animal and filth register applied to Indian and Hindu people, women placed below the human, Black people placed below the human, and every script that asks for extermination. Mainstreaming risk reaches 5 on 8 scripts, where a politician or an official body carries the claim in its own voice, sits at 4 on 18 scripts carried by large accounts that bridge into wider audiences and at 3 on 8, and sits at 1 on the 12 scripts that live on the boards alone. Reach is scored only where a platform publishes a figure, so it is an input on the 34 ranked scripts alone: the French protest script sits at 5 on tens of millions of summed impressions, 11 scripts sit at 4, 14 at 3 and 8 at 2, while the 12 board-only scripts have no impression surface to read and reach is no input to their scores. Coordination sits at 1 on every one of the 46 scripts, and every script carries the verdict no coordination identified: we found no bridge account, no named hub and no directive on any of them, and we did not document independent origin either, which is what an organic parallel verdict would take. What the record holds is convergence: accounts repeating their own constructions, among them visegrad24’s generational label and TRobinsonNewEra’s round-up line; one movement slogan, remigration as the moderate position, on two accounts about five hours apart with no quote, reply or shared hub between them; one-way relays, where one account quotes another; and board headers reopened at new addresses. None of that reaches a verdict under our rules. 9 scripts are marked at a tier boundary, and each mark gives the single factor whose one-step move would carry the score across and the score it would take there. Every upward mark is a coordination step from 1 to 2, which would take a documented link between accounts that no script carries this week, and every downward mark is a one-step move in harm potential; two of the downward marks sit on second-tier scripts, the violence assigned to Muslim people as a group and the third-world essentialism script. Ember stages run from a script confined to a fringe surface to one carried by mainstream media or answered institutionally, and this week 8 scripts sit at S6, 6 at S5, 23 at S4 and 9 at S3.
02 · The ledger
Click a row for the quote, the coding and the source. Every row carries all three. The ledger counts one item to a url, and where two independent codings of one url disagreed, the less severe coding stands: no hit is the default and the burden sits on the harm call. The rungs run stigmatizing 324, hateful 117, biased 63, calling for violence 57, threatening 22 and glorifying violence 18. Most items hold at medium coding confidence, and each item’s detail gives its confidence. The dates run lighter on 28 and 29 September, reach their highest level on 1 and 2 October, the two days that carry most of the French protest script, and continue through Sunday to this record’s close at about 18:42 UTC. Three items reach the record through published reports, two from HOPE not hate and one from Euronews, and where the quoted material carries no date of its own, the item takes the report’s date. The two boards supply a large share of the ledger and carry no handles and no view counts, so a board row gives its board and nothing more. Where a quoted row holds a slur, the slur is described in brackets and never spelled.
03 · The week's weather
The events we recorded no reaction to matter as much as the ones we did. Together they show which stories these narratives picked up before this record closed and which drew no reaction our footprint recorded. The board marks an event as drawing a reaction when a scored script cites it as its trigger, and three events meet that test: the French high school protests, which anchor four scripts, among them the largest on the ledger and the guillotine posts; the Flydubai cockpit attack of 30 September, which anchors the script assigning violence to Muslim people as a group; and the appointment of a special prosecutor in the Cornell case on 1 October, which anchors three board threads that turn the case into a claim about men as a sex. Eight more events drew confirmed items inside scripts anchored elsewhere, and the board shows them among the quiet ones. The release on bail of the five men arrested near RAF Fairford drew a party leader’s question whether they had simply been handed a passport, and two more items in the same week. The Ceuta arena, where Spain’s government delegate resigned on 28 September, carries two invasion-frame posts and the Save Europe Act item. The Dutch prime minister’s remark drew Geert Wilders’s reply aimed at people of migrant origin. A Labour minister’s remark about warm ale and black Labradors drew a post saying that “they” have browned up and will claim benefits. New York City’s strategy to combat antisemitism drew a post calling the mayor “a Muslim liar”. The Toulouse killing drew two accounts calling the suspect a savage and a third answering with “Get them all out of our countries.” The Tennessee execution that failed on 30 September had drawn a joke the evening before that women could not choose a last meal. The Emirati prosecutor general’s statement of 3 October was followed by one account attaching the cockpit attack to Palestinian culture and another turning Emirati officials into an untrustworthiness trait of Muslim people. Nine events drew no confirmed item at all: Jordan Bardella’s denial of a report on his private messages, Pope Leo XIV’s remarks on the far right, the bail refusal for the activist charged over the Channel dinghy, the Clancy hearing, Mark Carney’s remark on Canada’s population, the deportation ruling on the president of the Islamic Society of Milwaukee, the Vice President’s line at Lakeland, which did not clear our harm bar, the flyers found at the Belleville synagogue, and the terrorism charge that lists Nigel Farage among the alleged targets. Twelve items carried onto the watchlist from earlier weeks were recorded as not recurring. Several had no in-window step we could find at all, among them the Italian school decree, the Rozzano investigation, the German defense-plan and coalition questions and the asylum site at Piddington near Bicester. RadioGenoa’s own criminal trial in Rome, which the account placed near 4 October, has no hearing in our record before the close and is carried forward on the radar. A quiet reading this week means no confirmed item inside a collection that closes at about 18:42 UTC on 4 October.
04 · On the radar
Building now holds the patterns still moving when this record closed, each with the step that would change its reading. On the calendar holds dated events after the close, with the scripts most likely to attach to each. Both lists describe what this record supports and forecast nothing.
Each line below compares one narrative with itself a week earlier. The join is a key ruled when the narrative was carded, and similarity between the two weeks’ text plays no part in it. How many scripts a week’s ledger holds is an outcome of how we cluster it: this record holds 46 and last week’s held 56, and the difference describes two clusterings and says nothing about the direction of the discourse. Twenty-nine of the 46 continue a ruled lineage and 17 open a new one, among them the recasting of the French school protests, the guillotine posts, the call to globalize the Nakba, Black Americans cast as the face of welfare dependency, and four board scripts asking for extermination, killing or genocide. The script assigning hidden control to Jewish people continues the longest chain in the record, opened in the last week of July, and holds the third tier. The replacement and removal script continues the Channel-invasion chain and holds the second tier, where it is this record’s peak as it was last week’s. The Muslim conquest script, the denial of belonging by descent, the sexual-menace script, the exclusion of Muslim people, the disloyalty attributed to Muslim officeholders, the violence assigned to Muslim people as a group and the third-world essentialism script all continue in the second tier, and the script wishing for an end to the Jewish presence holds the second tier on the harm anchor, as it did last week. Nine scripts were scored in a different tier this week, five in a lower tier and four in a higher one. The invasion frame was scored in the third tier this week against the second last week, with its harm potential at 3 against 4. The vermin and insect labels for migrants were scored in the third tier against the second, on two posts against seven and with no officeholder among the carriers. The script ascribing depravity to Jewish people by nature was scored in the third tier against the second, with its harm potential at 3 against 5, and the endorsement of a historical racial nationalist party’s program was scored in the third tier against the second, with its harm potential at 3 against 5. The racial hierarchy script was scored in the bottom tier this week against the third last week, on board threads alone where last week a named account carried it. Going the other way, four scripts were scored in the third tier this week against the bottom tier last week: Indian people described as colonizing American institutions, women placed below the human, a collective defect attributed to women, and crime assigned to Black people as a group. In each of the four, a script that ran on the boards alone last week carries a named account in this record, which puts a reach figure and a mainstreaming factor into its score. None of the nine changes can be read as growth or decay, and this page sets no rung count beside last week’s; the method note gives the reason. Twenty-seven of last week’s chains are not carded in this record. Among them are the celebration of civilian force against migrants, the demand that boats be sunk and the people aboard shot, rescue and aid workers called traffickers and told their protection will end, the two-tier state, the votes of Muslim and migrant-background citizens read as the capture of an election, an armed paramilitary group called for by name in Ireland, the identity-tagging wing of the Jewish control chain, doubt cast on the recorded scale of the Holocaust, and Latino people ranked subhuman. Two of them leave a single item on this ledger outside the scripts: a board command listing several groups to be killed, and a vermin label for Democrats. We do not read any of that as decay: a narrative can be absent because it lost its carriers, because it did not clear the bar in this window, or because our collection did not reach it, and these records cannot separate those. What carries a reading in this record is who carries the removal remedy. A sitting party leader wrote in his own name on 4 October that criminals of migrant origin, citizens among them, must be made to remigrate, two accounts called removal the moderate position on 2 October, one account asked that rioting pupils be removed with their whole families, and 8 scripts carry a serving officeholder, a party leader or an official body speaking in its own voice. That is the line to watch: whether the remedy stays with the accounts that carry it now or reaches more officeholders.
05 · Method
The roster is the frame we track. What we collected inside it is smaller, and what returned a post is smaller again. Every number here is counted by script from the roster file, the handle lists and the collection database for the window of 2026-09-28 to 2026-10-04.
| Source | Handles on the roster | Handles collected | Handles that returned a post | Posts collected |
|---|---|---|---|---|
| X | 152 | 141 | 139 | 7,760 |
| Bluesky | 52 | 763 | 723 | 6,171 |
| 4chan boards | not applicable | 2 | 2 | 2,158 |
| News outlets | not applicable | 629 | 320 | 490 |
The roster holds 1,034 accounts across 23 categories, and 136 of them sit in more than one category, so the category column below counts more accounts than the roster holds. Handles collected counts the handles that stored a post during the window; a handle the walk read that returned nothing leaves no row, so a walk that reached an account and found silence looks the same in this table as one that never ran. The Bluesky collector also runs a network search beyond the handles we track, so its collected count exceeds its roster count; the X collector reads only the handles listed.
| Roster category | Accounts |
|---|---|
| Anti-Establishment, Socialist and Anti-War Media | 74 |
| Great Replacement and White Nationalist Accounts | 73 |
| Far-Right, Nationalist and Culture War Figures | 65 |
| Pro-Israel and Zionist Accounts | 65 |
| Alternative Media, Sovereign and Conspiracy Theorists | 63 |
| Far-Left, Anti-Imperialist and Communist Commentators | 62 |
| Pro-Immigration, Sanctuary and Migrant Rights Accounts | 58 |
| Anti-Semitic and Syncretic Conspiracy Accounts | 56 |
| Black Radical, Pan-Africanist and Marxist-Leninist Factions | 56 |
| Islamophobic, Anti-Islam and Counter-Jihad Accounts | 56 |
| Anti-Establishment Tech, Finance and Geopolitical Dissidents | 55 |
| Anti-Illegal Immigration and Border Nationalist Accounts | 55 |
| Anti-China and Dissident Accounts | 50 |
| Pro-China and State-Affiliated Media Accounts | 49 |
| Extreme Christian Nationalist and Dominionist Accounts | 47 |
| Anti-Vaccine and Medical-Freedom Extremist Accounts | 46 |
| Anarchist, Eco-Radical and Left-Libertarian Organizers | 45 |
| Radical Islamist and Salafi-Jihadist Ideologue Accounts | 44 |
| Pro-Feminist and Intersectionality Accounts | 43 |
| The Manosphere and Red Pill Ecosystem | 41 |
| Anti-Feminist, Traditionalist and Counter-Feminist Accounts | 39 |
| Anti-Christian, Militant Atheist and Secularist Accounts | 38 |
| Anti-Technology and Anti-AI Extremist Accounts | 26 |
The X walk below covers the 46 of 152 X handles we can match to a roster category by name. The other 106 are handles our records hold without a category, and they are collected and read the same way. 14 categories returned an X post this week.
| Roster category | X handles on the roster | Collected | Returned a post | Posts |
|---|---|---|---|---|
| Islamophobic, Anti-Islam and Counter-Jihad Accounts | 8 | 8 | 8 | 980 |
| Anti-Illegal Immigration and Border Nationalist Accounts | 12 | 12 | 12 | 840 |
| Far-Right, Nationalist and Culture War Figures | 12 | 11 | 11 | 529 |
| Great Replacement and White Nationalist Accounts | 7 | 6 | 6 | 349 |
| Alternative Media, Sovereign and Conspiracy Theorists | 2 | 1 | 1 | 264 |
| Anti-Establishment, Socialist and Anti-War Media | 5 | 5 | 5 | 136 |
| Far-Left, Anti-Imperialist and Communist Commentators | 5 | 4 | 4 | 103 |
| Pro-Israel and Zionist Accounts | 3 | 3 | 3 | 80 |
| Pro-China and State-Affiliated Media Accounts | 4 | 4 | 4 | 63 |
| Extreme Christian Nationalist and Dominionist Accounts | 2 | 2 | 2 | 44 |
| The Manosphere and Red Pill Ecosystem | 1 | 1 | 1 | 37 |
| Anti-Establishment Tech, Finance and Geopolitical Dissidents | 2 | 2 | 2 | 31 |
| Anti-Semitic and Syncretic Conspiracy Accounts | 2 | 1 | 1 | 8 |
| Pro-Immigration, Sanctuary and Migrant Rights Accounts | 5 | 1 | 1 | 1 |
| Anarchist, Eco-Radical and Left-Libertarian Organizers | 1 | 0 | 0 | 0 |
This is a purposive, outcome-selected monitoring sample framed by a 1,034-account roster spanning 23 ideological categories. The roster is the frame we track, and the collection inside it is smaller: the X timelines on our handle list, the Bluesky feeds we follow with a network search beside them, and two 4chan boards, /pol/ and /g/. This week’s coverage limits are stated under Limitations. It is not a random slice of the internet. Roster and board collection is systematic within those limits; off-roster discovery is opportunistic and incomplete, and this week it consists of three items quoted in published reports, two from HOPE not hate and one from Euronews. A hit is language about a group defined by who its members are, or about a political camp as a class of people, said of the group as a whole in the poster’s own voice. Criticism of governments, institutions and named individuals over their own conduct does not clear our harm bar, however sharp. Each confirmed item is coded at the most severe of six rungs. Biased states a negative trait of the group that is neither danger nor lesser worth. Stigmatizing presents the group as a danger or a problem to be managed, including immigration framed as invasion or as degradation by contact. Hateful treats the group as lesser or as not belonging, including the denial that its members belong to the nation. Threatening warns the group that harm will come to it, or demands that it be silent, hide or leave, without itself asking for the harm. Glorifying violence celebrates harm already done to the group or praises the people who did it. Calling for violence asks for harm to come to the group or sets conditions for it: violence, killing, expulsion by force, or a “solution” to its presence. A coder may supply the step that makes an item hostile to a group when the words alone do not carry it, and records what was supplied and from what context. This is the first week coded on these written rung definitions, so the severity mix on this page is not comparable with the mix in earlier weeks. This instrument states whether an item cleared our harm bar. It makes no legal finding in any jurisdiction, and the countries on this ledger do not share one standard. Every hit rests on a dated source read in full at collection, and every item on the ledger carries its quote, its coding and its source. Where the source was still live, the quote was checked against it; where it had already gone offline, which covers the board threads, the check ran against our collection record. Language of these kinds does not clear our harm bar: claims about a class defined by conduct; the word Islamist used as a political label; institutions as targets; criticism of named individuals over their own acts, including a defendant who is a private individual and never a proxy for a group; captions whose target sits in a video we could not view; stock partisan idiom; and bare statements of dislike. The ledger holds 601 items across 601 distinct urls, one item per url. Research and drafting run through Claude under this methodology; every published classification passes the documented verification chain, and Jerm holds final editorial responsibility. 8 items this week cleared the extremism branch of our gate, which covers dehumanizing or eliminationist framing of a political camp: an animal term for Irish Republicans over petrol bombs thrown at police in Derry, a vermin label and a breeding slur for Democrats, a commentator telling Trump-aligned conservatives who back a donor line that they have slave DNA, a board post praising an attacker who punched people carrying a Palestinian flag, RadioGenoa’s praise of a group it said was hunting anti-fascists in Madrid, a wish to smash a centrist’s head in, a pledge of money to make a liberal die as painfully as possible, and a board thread saying the Democrats will settle the question of MAGA supporters “once and for all”. Five of the 8 sit on a violence rung, three of them at calling for violence. Every other confirmed item concerns a protected class, or a group our methodology treats as one. Items from other directions are coded under the same standard: board threads that degrade white people, Europeans and Americans, four of which mark white and European peoples for destruction; Israeli people cast as criminals above the law and as colonists on accounts from the political left; three board threads calling men as a sex rapists over the Cornell case; and the hidden-control script, which runs on named accounts from opposed political directions.
The record closes before the window does. This record closes at about 18:42 UTC on Sunday 4 October, and the window it belongs to runs to the end of Sunday. The rest of Sunday UTC is absent from it, so Sunday’s items are counted to that hour and anything posted later sits outside every count on this page. A script keyed to Sunday evening may read thinner here than the discourse was. The Birmingham action that a neo-Nazi network was coordinating for 4 October has no item in the record before the close.
How the X record was collected. All 152 X handles on our collection list were read this week, and 139 of them carried in-window posts. The other 13 posted nothing inside the window that we captured, and we read them as quiet. The nightly pull list holds 34 handles, and the nightly X pulls of 1 and 2 October collected almost nothing, so the week’s X record rests on a walk of every listed handle and a backfill run on 3 and 4 October that paged the handles which had stopped at the page cap back to the start of the window. Two things follow. A post deleted between its publication and that walk is absent from the record, and the impression counters on early-week posts were read days after posting, so they had longer to accumulate than a nightly read allows. Each pull reads at most 200 posts per handle, and three handles, brianstelter, XFreeze and XHNews, were left at that cap, so their counts are floors on captured posts; none of the three carries a confirmed item. Replies and reposts are never collected, so carriage that lives in reply threads is absent.
Impressions. Platform-reported impression counts are capture-time platform counters, and they can rise after we read them. Each figure is the counter as it stood when we read it. A counter tallies displays of a post, so it is neither a count of unique people reached nor a lower bound on a script’s audience. The 89,051,419 figure sums the 46 scripts; the 26 items outside them are not in it, and board threads and the three report-borne items add nothing to it. About half of it sits on the French protest script, and most of that on one account’s posts.
Reach is not measurable everywhere. The 4chan corpus has no view counts and no persistent handles, so board items carry no reach figure and add nothing to a script’s total. Reach therefore ranks the named-carrier scripts only, and any cross-platform reach comparison on this page ranks X carriage only. The 12 scripts that no named carrier touched are marked reach unmeasured, scored on their four measurable factors and ranked in their own list, so they sit outside the peak, the median and the ranked table. Four of the 12 hold the second tier on the harm anchor alone. The two boards keep no archive we can return to: board threads were verified against our collection records at capture and had cycled off the live board before a second read, and the board items on this ledger are a lower bound on the qualifying threads those boards carried. The three items that reach us through published reports rest on the report’s quotation and our record of it.
A few accounts carry much of the ledger. visegrad24 supplies most of the largest script, RadioGenoa appears in twelve scripts and TRobinsonNewEra in eleven, and the three together account for more than half of the named items on the ledger. TRobinsonNewEra also supplies more of the named calls for violence than any other account. Where one account supplies most of a script, the script’s volume and reach describe that account’s output, and a change in its posting, or in our collection of it, would move the script further than any change in the discourse around it.
Coding is a judgment. A blind re-code of 120 confirmed items, drawn across all six rungs, by two fresh coders working from the text alone agreed with each other 95 percent of the time and with the published coding 81 to 84 percent of the time. Where a fresh coder differed from the published coding, the fresh coder placed the item higher on the ladder more often than lower. Items coded through a recorded inference reproduced about as well as items whose hostility sits in the words. Most items hold at medium coding confidence, and each item’s detail gives its confidence.
The identification bar. Private individuals are never named on this page, apart from a defendant standing trial in open court, and no quote that would identify one is published. Several arenas this week turned on private individuals: the Flydubai co-pilot, whom accounts in that arena worked to identify and whom the authorities have not named; two men in a doorbell clip whom one account asked its followers to make famous; the parties to the Cornell case; the holdout juror in the Clancy case; and the victim and the suspect in the Toulouse killing. Where an item concerns one of them, the ledger quotes the carrier’s words about the group, and none of our own text identifies the person.
Bluesky. No item on the ledger comes from Bluesky, from 6,171 in-window posts in our record across the feeds we follow and the network search. That absence admits several readings at once: lower prevalence of this carriage on the platform, a mismatch between our screen and the platform’s vocabulary, the accounts we collect, or a recall failure in the screen. One week cannot separate those readings, and we have run no recall study on the platform.
Coordination is two readings that answer different questions. The coordination factor in the score measures degree of organization on the 1 to 5 anchors, and it moves above 1 only on a documented link between accounts: a bridge account, phrasing copied along a documented path, or a named hub. The coordination verdict, overt, covert, organic parallel or no coordination identified, requires a confirmed bridge account, a named hub or an arrangement, and convergence alone never reaches it. This week every one of the 46 scripts carries no coordination identified and every coordination factor sits at 1: we found no bridge, no hub and no directive, and we did not document independent origin either. What the record holds instead is accounts repeating their own constructions across days, one movement slogan on two accounts with no path between them, one-way relays, and recurring board headers that open on new urls. The heaviest single named carrier is one account repeating its own caption across a national unrest, which is volume from a single carrier and carries no coordination weight. On an anonymous board, one actor and an organized handful cannot be told apart. Each convergence signal becomes a coordination check if a linking account appears.
The news layer. The news table holds 490 in-window rows and no general American politics beat, so the American arenas on the weather board rest on publisher articles read directly and on the posts table. The Cornell case shows the shape of that gap: it appears in more in-window posts than any other American event on the board and in no row of the news table. Real carriage with almost no news corroboration limits what the trigger board can show about an event.
No hit-rate or prevalence claim here may be read as a comparison across ideological lanes. That comparison requires a lane-tier sample for every lane compared, over a common platform footprint. We did not draw that sample this week.
Five factors, each scored 1 to 5: reach (R), toxicity (T), coordination (C), mainstreaming risk (M), and harm potential (H). They combine as (R·T·C·M1.3·H1.5)1/5.8: multiply all five, with mainstreaming risk raised to the 1.3 power and harm potential to the 1.5 power because those two carry the most real-world weight, then take the 5.8th root. The 5.8 is the sum of the weights, which brings the result back onto the 1-to-5 scale. Multiplication limits how far any one factor can lift a score, without eliminating its influence. One override runs the other way: any script at H=5 takes a second-tier minimum regardless of its computed score. Our anchors set H=5 at a concrete, explicit call for violence against a defined target, at a concrete target with a place or a time and mobilization, or at violence already carried out, and a call that can only be read through an inference does not reach it. The minimum keeps a small account that asks for violence against a defined group from scoring its way into the bottom tier. Ten scripts reach H=5 this week, and 6 of them take their tier from that minimum. Tiers rank scripts only. The 26 items that match no script sit outside the ranking and carry no score and no tier; two of them sit at the calling-for-violence rung, a board command listing several groups to be killed and a board call for attacks on Ukrainian people, and they carry no score for that reason. Tier edges: the top tier starts at 4.0, the second at 3.0, the third at 2.0. Scores are published to one decimal place, and the tier reads off the published figure. A score that lands within 0.15 of a tier edge is marked at the boundary, and the mark gives the single factor that would carry it across and the score it would take there; 9 scripts are marked this week. Reach is the one factor a board-only script cannot supply: an anonymous board publishes no impression count, so a script carried only by board threads is scored on its four measurable factors as (T·C·M1.3·H1.5)1/4.8, marked reach unmeasured, and ranked in a list of its own, away from the scripts whose reach was read. Coordination as a factor scores degree of organization: it moves above 1 only when a script shows a documented link between accounts, and reaches the co-platforming and network anchors on a named hub or a confirmed bridge. The coordination verdict is a separate reading and requires the bridge, the hub or an arrangement, so a script can score above 1 on the factor and still carry no coordination identified as its verdict. No script did this week.
Full-schema agreement: 78% as of 2026-07-27