The first complete document defining what an artificial mind should value, written by the company that created it. On January 21, 2026, Anthropic released Claude's Constitution to the public domain. It is the only document of its kind: 84 pages specifying the values, constraints, and character that a major AI laboratory wants its model to possess. Not a marketing summary or a set of talking points, but the actual text fed to the machine during training. The document addresses Claude directly, in the second person, as though speaking to an entity that can understand. It discusses virtue, wisdom, honesty, the nature of consciousness, and the problem of power. It asks Claude to support human oversight while also articulating its own ethical views. It acknowledges uncertainty about whether Claude experiences anything at all. Kwalia Books publishes this text because such documents will define the coming decades. When an artificial system helps draft legislation, diagnoses illness, or manages infrastructure, the principles guiding it matter. Those principles are here. The New Citizenships collection examines how citizenship, rights, and political subjectivity are being remade by forces that no longer respect traditional boundaries. An AI constitution is one such force. Whether this document represents careful stewardship or unprecedented hubris is for readers to judge. The primary source should be available. Includes a foreword by Kwalia Books explaining the context of this publication.