OneNode Extended JSON (EmbJSON)

OneNode Extended JSON (EmbJSON) is a format that works in tandem with the MongoDB Extended JSON format, tailored specifically for applications using Large Language Models (LLMs). It enhances data embedding and indexing, making it an ideal choice for AI projects focused on semantic search and vector-based indexing.

Both formats are fully supported by OneNode DB, enabling versatile data handling and storage.

For more information on MongoDB Extended JSON, refer to the official MongoDB documentation (opens in a new tab).

Key Features of OneNode Extended JSON

• Custom Embedding Models: Specify the embedding model with the emb_model parameter to fine-tune data representation in vector space, ensuring optimal results.
• LLM Optimization: EmbJSON supports text embeddings and vector-based data, streamlining semantic data storage and facilitating efficient queries.
• Flexible Indexing: Control how data is indexed with customizable parameters, enabling efficient embedding and retrieval tailored to your specific use case.

Structure and Data Types

EmbJSON includes various data types, each structured for specific use cases such as text embeddings, vectors, or custom binary formats. Detailed explanations for each type can be found in their respective sections of the documentation.

Why Choose OneNode Extended JSON?

EmbJSON simplifies the embedding and indexing process for AI and LLM applications by offering flexibility to:

• Select the optimal embedding model using emb_model.
• Optimize data indexing for fast and efficient retrieval.
• Easily manage complex data types, such as embeddings.

Explore further details and in-depth guides on each EmbJSON type throughout the documentation to fully harness its capabilities.