📖Architecture
Last updated
Last updated
ChatBees is an advanced Retrieval Augmented Generation platform provided as a self-managed service. ChatBees enables data storage, processing, and retrieval solutions that are faster, easier to use, and far more flexible than traditional offerings.
ChatBees combines a completely new RAG engine with an innovative architecture natively designed for the cloud. ChatBees uses a central data repository, the cloud storage, for persisted data that is accessible from all compute nodes in the platform.
ChatBees unique architecture consists of two key layers:
When data is loaded into ChatBees, ChatBees parses the data, splits data into chunks, generates the embedding for each chunk, and persists all data into cloud storage. ChatBees reorganizes that data into its internal optimized and compressed format.
ChatBees manages all aspects of how this data is stored — the organization, compression, metadata, statistics, and other aspects of data storage are handled by ChatBees. The data objects are not directly visible nor accessible by customers; they are only accessible through provided APIs.
The elastic service layer is a collection of services that coordinate activities. These services tie together all of the different components of ChatBees in order to process user requests. The cloud services layer runs on compute instances provisioned by ChatBees from the cloud provider.
Services managed in this layer include:
Authentication
Access control
RAG pipeline
Data and Vector Index
Security: encryption, secret key management
Private LLMs - in the future
