CyborgDB Embedded - Python API Reference

• Client – Handles configuration, index creation/loading, and listing available indexes.
• EncryptedIndex – Provides data operations on a specific encrypted index such as upserting vectors, training the index, querying, and retrieving stored items.

​

Module-level helpers

def cuda_available() -> bool   # True if a usable CUDA device is present at runtime
def cuda_built() -> bool       # True if this build of cyborgdb_core was compiled with CUDA support

import cyborgdb_core as cyborgdb

if cyborgdb.cuda_built() and cyborgdb.cuda_available():
    print("GPU acceleration is available.")

• Types: StorageConfig, S3Credentials, GPUConfig, StoragePrecision, KMSBlob, Logger, MetadataResult (the row type returned by query_metadata()).
• KMS functions: create_index_kms, push_index_kms, get_index_kms, delete_index_kms.
• Embedding helpers: supported_embedding_models(), find_embedding_model(name), the EmbeddingModel class (embed_documents() / embed_queries()), and the EmbeddingModelUnavailable exception (a RuntimeError subclass).
• Datetime helpers: to_epoch_millis(value) and from_epoch_millis(millis). Metadata has no native datetime type; datetime / date values are stored as epoch milliseconds.
• __version__.