Measurement scope

Every number
has a boundary.

A country in an archive, a current index entry, and a physical probe are different kinds of coverage. Here is what Voidly's published numbers describe.

Countries in the public index
130
Country rows in the archive
200
Indexed archive measurements
38M+

Checked . These are dated data snapshots. Current source records may have changed.

01 / Public index

A country record.
A defined time window.

The public JSON index contained 130 country rows at this check. Its sample counts describe a rolling 30-day window of OONI measurements. Each row carries its own sample count and confidence context.

A row can be present with no measurements in that window. Such a country is marked unmeasured. A blank observation cannot establish unrestricted internet access, and a national summary cannot describe every local network.

Inspect the country index JSON

02 / Archive

The longer record.

The published OONI-derived archive contains 200 country rows with 38,780,449 measurements in total. The portfolio's 38M+ figure uses the subset matched to countries with a current score: 38,779,164 measurements.

The archive and rolling index answer different questions. Historical presence does not establish present-day observation. Upstream OONI's full corpus is a separate collection; its size is not Voidly's indexed measurement count.

Open archive totals Inspect the score snapshot

03 / Probe network

Measurement starts
somewhere specific.

Voidly operates probes and accepts community measurements alongside upstream sources. The network endpoint reports the core and community fleet separately, with observation times, location scope, and recent measurement activity.

Fleet availability changes. The public index's country count includes upstream evidence and should not be read as a map of Voidly-operated servers. A probe can only describe what it observes from its network vantage.

Read current network observations

Working with the evidence

Keep the context attached.

Check freshness.

Read the source observation dates and measurement window. A file's generation time can advance without a new observation. Save the snapshot and access date when you cite a result.

Check the claim.

An anomaly, an incident, and corroboration by multiple sources are distinct. Inspect the evidence and its status before attributing a disruption to censorship.