Workflows

When you upload a document to Ragie, it goes through a series of processing steps before it becomes searchable and usable for retrieval. At a high level, these steps are

1. Extraction / Parse - The document is analyzed and converted into structured elements like text, tables, and images
2. Chunking - The extracted content is split into smaller segments optimized for retrieval
3. Indexing - The chunks are embedded and stored so they can be retrieved by queries.

By default, each document goes through all of these steps.

Workflows allow you to stop processing at an earlier step, depending on how you plan to use the document.

Currently, Ragie supports stopping the workflow at:

• Parse
• Index

Parse

If you choose "Parse", processing stops after the extraction step.

This means:

• The document is parsed into structured elements - available via the Elements API
• No chunking is performed
• No indexing occurs
• The document will not be available for retrieval queries

This workflow is useful when you want to

• Inspect or process the raw extracted document structure
• Perform custom chunking or indexing outside of Ragie
• Use Ragie primarily as a document parsing layer

Index

If you choose "Index", the document goes through all of the steps

• Parse
• Chunk
• Index

After processing completes

• Parsed elements are available through the Elements API
• Chunks are available
• The document can be used in retrieval queries

This is the default workflow and is recommended for most cases.

Summary