Every reader who encounters the subject of the Uyghurs eventually meets the same objection: how do we really know? The region is closed to independent journalists. The Chinese government calls the reporting lies. There are no press conferences in the camps. It is a fair question, and this archive takes it seriously, because the answer is one of the remarkable stories of modern documentation. What happened in the Uyghur homeland is not known from a single source, or a single kind of source. It is known from the convergence of at least four: the state’s own documents, the view from space, the testimony of survivors, and the painstaking work of researchers who tied these together. Each has limits. Together they form one of the most thoroughly documented human rights crises of this century.
The State’s Own Words
The most damning evidence did not come from the Chinese government’s critics. It came from inside the system. In November 2019, the New York Times published more than four hundred pages of internal Communist Party documents, now known as the Xinjiang Papers, including speeches in which the leadership ordered officials to act against religious extremism with “absolutely no mercy.” Days later, the International Consortium of Investigative Journalists released the China Cables, an operations manual for running the camps: how to prevent escapes, how to enforce discipline, how to score detainees with points before any release could be considered. The document that ordered the doors double locked was written by the region’s own security chief.
More leaks followed. The Karakax List, a spreadsheet from a single county, recorded hundreds of detainees along with the reasons for their internment: growing a beard, applying for a passport, having too many children, having a relative abroad. In 2022, the Xinjiang Police Files brought something the earlier leaks had not: faces. Thousands of photographs of detainees, hacked from police computers, alongside internal shoot-to-kill orders for escape attempts, published by a consortium of international media. These are not activist claims about the system. They are the system describing itself.
The View from Above
A government can deny documents; it cannot easily hide buildings. Researchers, most systematically at the Australian Strategic Policy Institute, used satellite imagery to identify and track more than three hundred and eighty detention facilities across the region, watching them grow walls, watchtowers, and factory wings year by year. Journalists at BuzzFeed News, whose reporting won a Pulitzer Prize, cross-referenced satellite images with a censorship quirk in Chinese mapping software to locate camps, then confirmed them with former detainees who recognized the layouts. The same methods documented the destruction of mosques and shrines described elsewhere in this archive. Satellite evidence has a particular value: it is independent, repeatable, and time stamped. Anyone with the coordinates can look for themselves.
The Voices
Documents and imagery establish the scale; only people can say what happened inside. Hundreds of former detainees and camp employees have now testified publicly, from Kazakhstan, Turkey, Europe, and North America. Their accounts, given separately and across years, describe the same intake procedures, the same political indoctrination sessions, the same punishments, the same crowded cells. Consistency of independent testimony is itself a form of evidence, and it is treated as such by courts everywhere.
Alongside the survivors stands the diaspora’s own record keeping. The Xinjiang Victims Database, a volunteer project, has collected more than one hundred thousand individual entries, each naming a detained or disappeared person, many submitted with documentation by relatives willing to be identified. This matters because it converts a statistic into a list of names, and lists of names can be checked, challenged, and corrected. Researchers such as Adrian Zenz mined the government’s own statistical yearbooks and procurement notices to document the collapse in Uyghur birth rates and the spending on camp construction, findings later scrutinized and substantially confirmed by other scholars and news agencies.
The Verdicts
Institutions that weigh evidence for a living have now done so. The United Nations human rights office, after years of delay and under heavy Chinese pressure, concluded in August 2022 that the extent of arbitrary detention, together with the restrictions on religious and cultural life, may constitute crimes against humanity. The Uyghur Tribunal in London, an independent panel that heard witnesses under cross-examination and published its evidence, found in 2021 that genocide had been committed through the suppression of births. Parliaments in several countries, including Canada, the Netherlands, France, and the United Kingdom’s House of Commons, have passed resolutions describing the policies as genocide or crimes against humanity, and two successive United States administrations, of both parties, formally designated them genocide. One can debate the legal terms. What no serious institution that has examined the record disputes is the underlying facts.
What We Cannot Know
An honest archive also states the limits. No one knows the exact number of people who passed through the camps; the widely cited figure of a million or more is an estimate built from leaked ratios, satellite capacity, and official data, not a headcount. No one knows precisely how many facilities still operate, or how many detainees were transferred into prisons through the mass trials of 2017 to 2019, though sentencing statistics published by the government itself show prison terms in the hundreds of thousands. Independent verification inside the region remains impossible; the guided tours offered to selected visitors are not that. These uncertainties are real, and this archive will always name them. But uncertainty about the precise number of victims is not uncertainty about whether there are victims. We do not know everything. What we know is enough.
That is why this archive exists: not to add new claims, but to hold the existing record where readers can find it, follow every link to the original source, and judge for themselves.