How it works

Three things you can do with Newjee.

Newjee turns any query — a keyword, a URL, a topic, an entity, a breaking event — into a structured analysis. Three primary surfaces, each calibrated to a different question.

What you type

The query bar accepts five things

A URL
https://reuters.com/world/article-123
Reality Card on that single article — claims, source health, framing, manipulation indicators.
A keyword
hormuz closure
Live phenomenon view. Newjee finds recent coverage, groups it into editorial blocs, surfaces the divergence.
A topic
EU AI Act enforcement
Phenomenon scoped as ongoing. Auto-extends the time window as new signals land.
An entity
Vladimir Putin
Entity-centric view. Reality Field shows narrative clusters, Trust spectrum maps sources by reliability.
A state of reality
Oil & supply shock Q2 2026
Crisis-class phenomenon. Snapshots every 6h; perturbations flag narrative drift > 2σ over baseline.
A publisher / journalist
reuters.com or @glenngreenwald
Media Entity Profile. Domain-decomposed trust, bias radar, framing footprint, audit history.

What you get back

Reality Card — single-article analysis

Atomic unit. URL or keyword in, structured card out: claim status, source reliability, framing, manipulation indicators (with verbatim evidence spans), propagation severity. ~15 seconds.

Try a Reality Card

Reality Field — navigate the epistemic landscape

Fullscreen canvas. Soft cluster blobs colored by epistemic kind (fact / interpretation / propaganda / unknown), sized by volume, opacity = confidence. Layer toggles, trust spectrum, timeline scrub.

Open the Reality Field

Phenomena — comparative bloc analysis

The operator workbench. Density matrices per editorial bloc, pairwise divergence with the Spectral Lens (rank-2 frame + manipulation axes), inferential validation (permutation null + bootstrap CI), narrative timeline.

See phenomena

Media intelligence — outlets, journalists, opinion leaders

Trust scores per actor, domain-decomposed (politics / economics / science / tech / culture / war / health / climate). Bias radar, trust evolution. Any actor can contest its score.

Browse media

Forks & discussion — analyses are social objects

Phenomena can be forked GitHub-style. Plain-text discussion at the bottom of every thread. "Useful" flag is binary — no upvotes, no popularity contest.

See phenomena

Browser Kit — Reality-check anywhere

Chrome / Edge / Brave extension. Click on any article, post, or transcript — get the Reality Card inline, no copy/paste.

Get the Browser Kit

How they connect

The same engine drives every surface. A Reality Card is one article's epistemic vector. A phenomenon stacks many such vectors per editorial bloc into density matrices. A media profile aggregates a publisher's vectors across all phenomena. The Browser Kit gives you the same Reality Card wherever you read.

Result: every claim you see in Newjee can be traced backward (which signals support it?), forward (how is it propagating?), and laterally (how do other blocs frame it?).

Pick one and try it

Reality Card (URL or keyword)Reality Field Media intelligence

For the math behind it, see methodology or the published papers.