Documentation

Glossary

Definitions of the Spark terms used across the docs.

Terms

The Spark terms used across the docs.

Bucket: A container for objects. A smart bucket indexes its objects for search and query.
Object: A file plus its key. Spark accepts any file type.
Smart bucket: A bucket that Spark reads and indexes as you write to it.
Tenant partition: A slice of a bucket isolated by your own customer, set with the X-Spark-Tenant header.
Agentic query: A natural-language question that Spark answers with citations.
Effort: How hard the query agent works. Standard is fast. Ultra searches more thoroughly.
Workspace: The container for buckets, keys, and members. The unit of access.
Answer engine: A query mode that returns a natural-language answer with sources.