Loading the page…
This takes a few seconds the first time; after that it is cached.
Step 1 of 2 · Analyzing the conversation0 %
Starting…

THEVERIFICATOR.NET Lite

Español

Detect influence operations by the shape of the network, not by volume.

This tool detected some 500 automated accounts that amplified CJNG cartel disinformation during the road blockades after the death of “El Mencho”. It was picked up by AFP, OCCRP, Milenio and La Opinión. Now you can do the same with your topic.

Try:26-S housing protestspemex / huachicol“El Mencho” case
How to write a search

Boolean: AND/Y, OR/O, NOT/NO or -word, "exact phrase", parentheses and * for prefix. Example: (pemex OR huachicol*) AND -gasolina

Period: 29/09/2026 – 13/10/2026 · the whole captured conversation

Amplification network

How to read the graph

Each dot is an account; its size is how much amplification it receives and its color is the category the classifier assigns. Lines show who retweets or mentions whom. Communities pull apart on their own because of how each group operates (ForceAtlas2, the same algorithm as Gephi). Height is time: accounts that appear later sit higher. Drag to rotate the 3D graph and zoom with the wheel. The public version shows the names of the 60 most amplified accounts; the full network, with every account and its figures, comes in the CSV.

Loading the 3D engine… (the first time takes a few seconds)
≈97.1 %
account diversity: distinct accounts per 100 posts (lower = few accounts post a lot)
70.0 %
are retweets (amplification)
0.2 %
of the volume comes from the 10 most active accounts
0
burst days (more than 2 deviations above average)
12/10
peak day: 10.80 % of the whole period in a single day
Weak signals

Automated amplification reading. It is a guide to decide where to look, not a verdict.

No strong signals of concentrated amplification in this range.

Activity classification

Each post is classified using seven behavioral criteria of the account. It is a presumptive reading: signals, not conclusions.

Organic: 88.17 %Coordinated: 9.34 %Artificial: 2.49 %

Accounts most mentioned or retweeted by the activity of each category. The % is their share of all mentions and retweets made from that category.

Top 10 most mentioned · organic

@petrogustavo 🔍0.55 %
@isaacrrr7 🔍0.47 %
@ericldaugh 🔍0.44 %
@fonsiloaiza 🔍0.39 %
@javiernegre10 🔍0.39 %
@cartoonsotmoon 🔍0.34 %
@abdelaespriella 🔍0.34 %
@pjmiesviu 🔍0.32 %
@pabloharour 🔍0.30 %
@revistasemana 🔍0.29 %

Top 10 most mentioned · coordinated

@claudiashein 🔍1.14 %
@petrogustavo 🔍0.79 %
@miyarah_ 🔍0.75 %
@andres_pinot 🔍0.71 %
@ricardobsalinas 🔍0.71 %
@azucenau 🔍0.68 %
@javiernegre10 🔍0.63 %
@jorgegogdl 🔍0.59 %
@mlopezsanmartin 🔍0.55 %
@aztecanoticias 🔍0.54 %

Top 10 most mentioned · artificial

@laderechadiario 🔍0.97 %
@petrogustavo 🔍0.97 %
@abdelaespriella 🔍0.90 %
@julianmaciast 🔍0.76 %
@pabloharour 🔍0.63 %
@eduardomenoni 🔍0.59 %
@ericldaugh 🔍0.56 %
@gunthereagleman 🔍0.45 %
@aztecanoticias 🔍0.45 %
@capitana_espana 🔍0.42 %

Daily activity

5.4 %10.8 %2026-09-29 · 6.89 % of the period29/092026-09-30 · 7.37 % of the period2026-10-01 · 7.37 % of the period01/102026-10-02 · 7.38 % of the period2026-10-03 · 7.38 % of the period03/102026-10-04 · 4.60 % of the period2026-10-05 · 3.69 % of the period05/102026-10-06 · 6.90 % of the period2026-10-07 · 7.39 % of the period07/102026-10-08 · 7.38 % of the period2026-10-09 · 7.38 % of the period09/102026-10-10 · 7.39 % of the period2026-10-11 · 7.38 % of the period11/102026-10-12 · 10.80 % of the period2026-10-13 · 0.70 % of the period13/10
■ activity■ bursteach bar: % of all posts in the period on that day

Most active accounts

@grok 🔍0.14 %
@silvestrica11 🔍<0.01 %
@newonpolsia 🔍<0.01 %
@unkn0wnski 🔍<0.01 %
@pilarro07192675 🔍<0.01 %
@normilgodoy 🔍<0.01 %
@montesforever 🔍<0.01 %
@monansnt 🔍<0.01 %
@michel778899 🔍<0.01 %
@maritafernande7 🔍<0.01 %

Most amplified accounts

@petrogustavo 🔍0.20 %
@isaacrrr7 🔍0.16 %
@javiernegre10 🔍0.14 %
@laderechadiario 🔍0.12 %
@miyarah_ 🔍0.12 %
@ericldaugh 🔍0.11 %
@revistasemana 🔍0.11 %
@pabloharour 🔍0.11 %
@fonsiloaiza 🔍0.10 %
@julianmaciast 🔍0.09 %

Hashtags (sample)

#lacasadelosfamososmx 🔍0.80 %
#nsfw 🔍0.60 %
#lacasadelosfamososméxico 🔍0.50 %
#nsfwtwt 🔍0.46 %
#diorsummer27 🔍0.44 %
#urgente 🔍0.41 %
#pfw 🔍0.41 %
#dior 🔍0.39 %
#linglingkwong 🔍0.36 %
#candid 🔍0.36 %
#lagranjavip 🔍0.34 %
#brosis 🔍0.33 %

Who retweets whom

@jgg19xd → @cathy_nwn<0.01 %
@jorge113o → @joseph_200k<0.01 %
@s_b_c_1987 → @indomableyletal<0.01 %
@surajgu46358933 → @bbyssweetgirl<0.01 %
@00koom1127 → @ormkornvn<0.01 %
@01c4t → @winnie62123994<0.01 %
@01johnsilver → @espartanoestoic<0.01 %
@01khrist → @gantielsigol<0.01 %
@0530daniel → @france24_es<0.01 %
@0800naufragio → @ladymoskigna<0.01 %
@0809_fernandez → @bblueyeee<0.01 %
@08mo01mo08 → @afycsopinof<0.01 %
@0_booom_6996 → @podemoslpgc<0.01 %
@0becky2i_m_back → @rabbijeffx<0.01 %
@0fffgunnn → @j0nasl4rry<0.01 %

Languages

es54.79 %
en44.71 %

Real cases · see them in 3D

SIGNAL · 26-S

The 26-S wave started on September 23, three days earlier

Housing and encampments: coordinated accounts show up before organic ones (p=0.006) and the non-organic share drops as volume rises. @eldiarioes, second most amplified account.

1,158posts
25.6 %non-organic
528communities
06/09 00:00 · 1107/09 00:00 · 1708/09 00:00 · 709/09 00:00 · 1910/09 00:00 · 2311/09 00:00 · 1012/09 00:00 · 1113/09 00:00 · 1814/09 00:00 · 1015/09 00:00 · 1516/09 00:00 · 1917/09 00:00 · 718/09 00:00 · 1619/09 00:00 · 1520/09 00:00 · 2221/09 00:00 · 1722/09 00:00 · 1523/09 00:00 · 5224/09 00:00 · 6225/09 00:00 · 4126/09 00:00 · 4227/09 00:00 · 5528/09 00:00 · 8229/09 00:00 · 9430/09 00:00 · 8901/10 00:00 · 5702/10 00:00 · 9503/10 00:00 · 8604/10 00:00 · 8405/10 00:00 · 67
See the case with its 3D graph → ARCHIVE · EL MENCHO

The night of February 22: from almost nothing to 329 posts per hour

The analysis picked up by AFP, OCCRP, La Opinión and Milenio: a highly concentrated core and 42.5 % non-organic activity.

15,225posts
42.5 %non-organic
246communities
22/02 04:00 · 122/02 11:00 · 1622/02 12:00 · 4122/02 13:00 · 5222/02 14:00 · 14922/02 15:00 · 30222/02 16:00 · 27622/02 17:00 · 28222/02 18:00 · 27522/02 19:00 · 26722/02 20:00 · 26122/02 21:00 · 26822/02 22:00 · 27222/02 23:00 · 26723/02 00:00 · 28623/02 01:00 · 28523/02 02:00 · 32923/02 03:00 · 30523/02 04:00 · 31623/02 05:00 · 31923/02 06:00 · 31223/02 07:00 · 31323/02 08:00 · 30223/02 09:00 · 32223/02 10:00 · 31623/02 11:00 · 30423/02 12:00 · 30323/02 13:00 · 28623/02 14:00 · 29523/02 15:00 · 28723/02 16:00 · 27623/02 17:00 · 28823/02 18:00 · 26623/02 19:00 · 24923/02 20:00 · 25423/02 21:00 · 26823/02 22:00 · 25323/02 23:00 · 25624/02 00:00 · 25024/02 01:00 · 24624/02 02:00 · 24524/02 03:00 · 25224/02 04:00 · 27224/02 05:00 · 26124/02 06:00 · 24224/02 07:00 · 25424/02 08:00 · 26024/02 09:00 · 25824/02 10:00 · 27624/02 11:00 · 25424/02 12:00 · 27824/02 13:00 · 25024/02 14:00 · 19024/02 15:00 · 18624/02 16:00 · 14524/02 17:00 · 18624/02 18:00 · 17424/02 19:00 · 19324/02 20:00 · 20824/02 21:00 · 19424/02 22:00 · 20824/02 23:00 · 224
See the case with its 3D graph →

🔓 Every analysis is published under Recent analyses, like a pastebin. With full access they are private.

Recent analyses

How it works

What it measures and what you get
  • 90 days of X/Twitter conversation you can explore.
  • 7 behavioral criteria separate organic from coordinated and artificial activity.
  • Amplification network of up to 3,000 accounts, with communities and bridge accounts, in ForceAtlas2.
  • Bursts, concentration and most amplified accounts, to decide where to look first.
How the analysis is rebuilt

THEVERIFICATOR.NET is a forensic analysis tool for influence operations on social media, built by journalist Alberto Escorcia (@AlbertoEscorcia). It detects coordinated inauthentic behavior from how the conversation moves: who posts, who amplifies, at what pace and in which groups.

  1. Capture. Public X/Twitter posts grouped by topic, country and language, plus external signals (Google Trends, GDELT, news RSS).
  2. Ingestion. Everything is normalized into an SQLite database with a full-text index; duplicate posts are resolved by canonical ID.
  3. Classification. Seven behavioral criteria separate organic, coordinated and artificial activity.
  4. Analysis. Time bursts, amplification graph (retweets and mentions), communities and bridge accounts.
  5. Publication. Every day the last 90 days are exported in monthly files and uploaded to this server.

The code, the full methodology and the backups live in the project’s private repository, documented so it can be rebuilt.

What this Lite version includes

Public viewer of the last 90 days of X/Twitter conversation with aggregates, activity bursts and the amplification network. No API and no post listings; the CSV (US$ 50) and the top 100 (US$ 100) are paid one-time downloads. Every search is published under “Recent analyses”; with full access (US$ 100 per month) your space is private.

Full version of THEVERIFICATOR.NET

Choose what you need. You pay with PayPal: the CSV and the top 100 are one-time downloads you get instantly; access gives you a code to use the site for a month.

Full accessUS$ 100for one month · 3 free downloads · your history and your own space, off the public list
Analysis by AlbertoUS$ 250per analysis · editorial reading of your topic
Download CSVUS$ 50one-time download, one row per account · access not included · «the whole conversation» 14 d
Top 100US$ 100one-time download · the hundred most amplified accounts · «the whole conversation» 14 d
▶ See the demo and samples of each deliverable

In the press

Screenshot of the article in AFP (via France 24)AFP (via France 24)Screenshot of the article in OCCRPOCCRPScreenshot of the article in La Opinión (EE. UU.)La Opinión (EE. UU.)Screenshot of the article in AFP (via Excélsior)AFP (via Excélsior)Screenshot of the article in AFP (via Animal Político)AFP (via Animal Político)Screenshot of the article in InfobaeInfobaeScreenshot of the article in MilenioMilenioScreenshot of the article in El CongresistaEl Congresista

Or write to us

In the press · articles that picked up the analysis

Articles on the disinformation that followed reports of the death of “El Mencho” that cite Alberto Escorcia’s analysis. Screenshots of the passage, linked to each original; only articles verified by opening them are included.

Who is behind it

Alberto Escorcia is a journalist who has investigated online disinformation since 2009. He worked with BuzzFeed and LoQueSigue, and runs THEVERIFICATOR.NET.

Its author has produced analyses for platform teams at Twitter, Google and Facebook, who checked them against their own tools and reached the same conclusions.
Muck Rack profileLinkedIn · work and references

Hire me

Method and limits

The tool measures with fixed, reproducible rules: the same search over the same period gives the same result, and every figure can be audited.

What it finds are behavioral signals, not a verdict on who is behind an account: they tell you where to look and what to verify. Before publishing a finding, check it against other sources, as was done in the El Mencho case.