The tradition's literature is layered, and the dating of the layers has been one of the most consequential arguments in Jewish scholarship.
The earliest strand is not called Kabbalah at all: Merkabah and Hekhalot mysticism from late antiquity, concerned with ascent through heavenly palaces to the chariot-throne of Ezekiel's vision. The Sefer Yetzirah, the Book of Formation, is a very short and very strange text on the ten numbers and twenty-two letters through which the world is made; its date is genuinely uncertain, with estimates spanning several centuries.
Kabbalah proper appears in Provence and Spain in the twelfth and thirteenth centuries, with the Bahir and then the Zohar, which is its central book. The Zohar presents itself as the teaching of the second-century sage Shimon bar Yochai, recovered after a thousand years. Gershom Scholem, who founded the modern academic study of the subject, established on linguistic and historical grounds that it was substantially composed in thirteenth-century Castile, most probably by Moses de Leon. That conclusion is mainstream in scholarship and rejected by much of traditional Kabbalistic Judaism, which continues to hold the ancient attribution. Both positions should be reported; they are not reconcilable.
Access was traditionally restricted, by a rule of thumb requiring the student to be male, married, forty years old and already accomplished in Talmud, a restriction meant to keep speculative material away from the unprepared, and one that modern Kabbalah-teaching organisations have discarded entirely.
Two later developments carried it outward. Hasidism from the eighteenth century made Kabbalistic ideas the devotional property of ordinary Jews. And Renaissance Christian scholars produced a Christian Cabala, which read the sephiroth as prefiguring the Trinity and the incarnation.