• Documents live in Collections (roughly: rows in a Table, in SQL terms)
• No fixed schema — documents in the same collection can have different fields
• Built for horizontal scaling (sharding) and high write throughput
• Query language is a rich, JSON-based syntax rather than SQL

{
  "_id": "64f1...",
  "name": "Ada Lovelace",
  "skills": ["math", "programming"],
  "address": { "city": "London" }
}