Hybrid vector database

A system comprising:

• one or more processors;
• and
• memory storing instructions that, when executed by the one or more processors, cause the system to:
• obtain vector data to be inserted into a vector database;
• index the vector data into a mutable index of the vector database by generating one or more key-value pairs for each vector of the vector data;
• periodically merge, in accordance with a first interval, the vector data from the mutable index into a first immutable index;
• periodically merge, in accordance with a second interval, the vector data from the first immutable index into a second immutable index;
• receive a query vector;
• and
• perform a vector search in the mutable index, the first immutable index, and the second immutable index based on the query vector.

One or more non-transitory computer-readable media storing instructions that, when executed by one or more processors, cause the one or more processors to perform operations including:

• obtaining vector data to be inserted into a vector database;
• indexing the vector data into a mutable index of the vector database by generating one or more key-value pairs for each vector of the vector data;
• periodically merging, in accordance with a first interval, the vector data from the mutable index into a first immutable index;
• periodically merging, in accordance with a second interval, the vector data from the first immutable index into a second immutable index;
• receiving a query vector;
• and
• performing a vector search in the mutable index, the first immutable index, and the second immutable index based on the query vector.