Learn about ChatBotKit's stores, an abstract storage class for efficient and organized storage and retrieval of data. Choose between Ada Sprout and Ada Loom depending on the size of your dataset.

ChatBotKit Stores are an abstract storage class that is used to retrieve information. They are a fundamental component that enables the efficient and organized storage and retrieval of data. Stores are exclusively used with Datasets. Upon creating the datasets, the user can select the appropriate store to be used to store the data. This makes it easy to manage different types of data across different datasets, providing an effective way to organize data and retrieve it efficiently.

There are two store classes that can be used with ChatBotKit: Ada Sprout and Ada Loom (default). Ada Sprout provides a lightweight storage option that can be used for small datasets, while Ada Loom is designed for larger datasets and provides a more robust storage option. By selecting the appropriate store, users can ensure that their data is being stored efficiently and effectively, allowing them to easily retrieve and analyze it as needed.

Ada Sprout

Ada Sprout is the first-generation store that combines text-embedding-ada-002 and a vector database. The store is ideal for smaller datasets where speed of search is critical. The technology behind this store results in rapid fetch times for small to medium dataset sizes.

Lingo Sprout

Lingo Sprout is second-generation store that combines text-embedding-3-small and a vector database. The store is ideal for any datasets suitable for semantic search. This store has a good density and it is generally good as a general-purpose store.

Ada Loom

Ada Loom is a second-generation store with a solid foundation that combines text-embedding-ada-002 and bm25. This store has been optimized for larger datasets and can handle millions of records seamlessly. Its technology base makes the store perfect for use-cases where highly accurate search results are required.