Savant provides native read and write support for the most common file types and file systems used in enterprise data workflows. This article describes the supported file types, their read/write limits, and the file system connectors available for sourcing and storing files.

Savant supports the following file types as both input sources and output targets in workflows.

Flat files are plain-text tabular formats, where each row is a record and each column is a field. Savant supports:

A text blob is a plain-text file that Savant reads as a single block of content rather than row by row. This is useful when working with unstructured or semi-structured text – such as exported logs, raw API responses, or large freeform documents – where treating the content as a table doesn't make sense.

Savant supports both Excel and Google Sheets formats:

Savant supports .pdf files for both read and write operations. PDF extraction capabilities depend on document structure; see PDF for details.

.parquet files are supported for columnar data workloads. Parquet is the recommended format for large-scale dataset processing in Savant.

The following table summarizes the read and write size limits by file type. Limits apply per file, per operation.

Savant integrates with all major enterprise file systems. Each connector is configured independently and supports its own authentication model, path conventions, and performance profile.

For setup instructions, authentication requirements, and connector-specific limits, refer to the individual connector guides.

Review the file type guides for detailed configuration options, parsing behavior, and known limitations: