The Maya wrote books, and almost none of them survive. Spanish clergy destroyed them systematically as idolatrous, most notoriously at Mani in Yucatan in 1562, and the climate destroyed most of the rest. Three bark-paper codices have been accepted for a long time, and a fourth was disputed for decades before analysis largely settled in its favour. That is the entire surviving library of a literate civilisation.
The man who directed the burning at Mani, Diego de Landa, also wrote the account that later made decipherment possible. His attempt to record the script as an alphabet was wrong about how it worked but preserved sound values, and it became the key that twentieth-century scholars turned. Yuri Knorozov argued in the 1950s that the script was partly phonetic, and Tatiana Proskouriakoff showed that the inscriptions recorded the reigns of actual rulers rather than only astronomy and ritual. Between them they turned the monuments into history.
Alongside the codices sit texts written in indigenous languages using the Latin alphabet after the conquest, of which the Popol Vuh is the most important, and Spanish accounts of Nahua religion compiled with indigenous informants.
Every one of those sources was produced under colonial conditions, by people with reason to shape what they said. The inscriptions carved before contact are the only body of evidence in the region that is not, and that is why their decipherment mattered so much.