How is the json that we submit to mongoDB is converted to bson?

When you submit JSON to MongoDB, you're actually working with a high-level representation that looks like JSON. Behind the scenes, the MongoDB driver for your programming language (Python, Java, Node.js, etc.) automatically converts that JSON-like structure into BSON (Binary JSON) before sending it to the database[citation:4]. BSON is a binary-encoded serialization format designed specifically for MongoDB's needs—it's faster to parse, supports more data types than JSON, and includes length prefixes that make it easier for MongoDB to jump between fields during queries[citation:4].

MongoDB uses BSON instead of JSON for three critical reasons. First, speed—BSON can be parsed much faster than JSON because of its binary structure[citation:4]. Second, rich type support—BSON adds non-JSON-native data types like Date, ObjectId, Decimal128, and Binary data that are essential for database operations. Third, traversability—BSON documents include length prefixes that allow MongoDB to skip over fields without parsing them, making queries significantly faster[citation:4]. These advantages come with a small trade-off: BSON documents are slightly larger in size than equivalent JSON due to the added type metadata[citation:5].

The conversion from JSON to BSON happens automatically in your MongoDB driver. For example, when you insert a document using a MongoDB driver, the driver serializes your data structure into BSON format. This involves encoding each field with its type identifier, field name, and value in a binary format. For nested documents and arrays, the same process applies recursively. When you retrieve data, the driver deserializes BSON back into the native data structures of your programming language[citation:9].

A BSON document is structured with binary elements that make it efficient to process. Each field is stored with a type identifier (1 byte), field name as a string, and the actual value in binary format[citation:4]. The document begins with a 4-byte integer indicating total document size. This length prefix allows parsers to skip entire subdocuments without parsing their contents—a key performance feature[citation:4]. Here's what a simple document like {"hello": "world"} looks like in BSON:

When you need to view BSON data in a human-readable format, MongoDB uses Extended JSON[citation:6][citation:8]. This is a string format that preserves BSON's type information using special wrapper objects. Extended JSON has two modes: Canonical mode emphasizes type preservation (e.g., {"$numberLong": "50"}), while Relaxed mode emphasizes readability (e.g., 50 for integers within safe range)[citation:6][citation:8]. Tools like bsondump and mongoexport use these formats to convert binary BSON files into readable JSON[citation:1].

For everyday development, you rarely need to think about BSON conversion—your driver handles it transparently[citation:4]. However, understanding BSON helps when debugging type issues, working with MongoDB tools, or building performance-sensitive applications[citation:4]. For example, when you query with db.collection.find(), the driver sends BSON to the server and receives BSON back, converting to your language's native types automatically. When using mongoexport to create JSON files, the tool converts BSON to Extended JSON so you can read the data[citation:1].