Documents, fields and the document model

MongoDB stores data as documents: records made of field and value pairs that look like JSON objects. A value can be a string, a number or a date, but also a nested document or an array (a list), so one customer document can hold its addresses inline.

Documents live in collections, and collections live in databases. Unlike a relational table, a collection does not force every document to have the same fields. That flexibility lets you shape each document around how the application reads it, often fetching everything a page needs in one query. It is also why data modelling matters so much later: nothing stops you storing a mess.