TROLL2 — Coordinated-Inauthentic-Behaviour Radar

Spot coordinated trolling before you engage.

TROLL2 helps you reason about whether a Facebook account is likely part of a troll farm or a coordinated-inauthentic-behaviour (CIB) campaign — using only publicly-visible signals. A locked profile hides its timeline from everyone, so we never see private data. This is decision-support, not surveillance: TROLL2 surfaces patterns; you make the call.

🧩

Install the browser extension

The extension is the primary way to use TROLL2. It runs in your browser and flags likely troll accounts inline as you browse Facebook, drawing on the same shared signals shown below.

Add toChrome Add toEdge

One-click install is coming — no Developer mode, no folders. Until then, load it from source (which also lets you audit exactly what it does):

Open chrome://extensions.
Turn on Developer mode (toggle, top-right).
Click Load unpacked and select the extension/ folder.

🔍

Public signals only

No private timelines, no scraping behind a login wall. Just what any visitor can see.

🕸

Coordination clusters

Accounts posting near-identical text in lockstep are grouped so patterns become visible.

⚖

ToS-aware & honest

A flag is a likelihood, not a verdict. We frame everything as support for your own judgement.

TROLL2 produces likelihood estimates from public signals. It is not proof of wrongdoing and should not be used to harass anyone. Always apply your own judgement.

Public stats

Aggregated, anonymous counts across everyone using TROLL2. No account names or personal data are shown here.

About TROLL2

TROLL2 analyses likely Facebook troll-farm and coordinated-inauthentic-behaviour accounts. Because a locked Facebook profile hides its timeline from everyone, the tool relies solely on publicly-visible signals — it never accesses private content.

🧭

Decision-support, not surveillance

TROLL2 highlights patterns associated with coordinated trolling so you can make an informed decision. A "troll flag" is a likelihood estimate, never a verdict.

What it looks at

Publicly-visible posting and reply patterns
Near-identical text shared across many accounts (coordination)
Bursty, lockstep activity timing
Other public-profile signals consistent with inauthentic behaviour

Troll-likelihood levels

Clean — no meaningful troll signal
Unlikely — minor signal, probably authentic
Possible — some signals worth noting
Likely — multiple converging signals
Clear — strong, consistent coordination signal

This tool provides likelihood estimates only. It is not affiliated with Facebook / Meta. Read our privacy policy.