Read your configuration

System is a read-only view of how your Space is configured. Open it from the user menu in the sidebar. No changes here — just transparency into how retrieval works. How the Space answers is Controls, not System.

General

Space ID, name, enabled status, and supported languages. Confirms which Space you are viewing.

Sources

Where content is ingested from — for example GitHub repositories, branches, and paths included in your Knowledge base. Shows what external connectors feed your Space.

Storage

Where vectors and documents are stored:

• Vector store — embeddings index (table, dimensions, index name)
• Document store — raw indexed content (bucket, table)

Assistant Model

The language model used to generate answers: model name, temperature, and base URL. This is what powers Chat responses.

Embeddings Model

The model that converts text into vectors for search. Dimensions should match your vector store configuration.

Pipeline

Chunking configuration — how documents are split:

• Chunk size — characters per segment
• Chunk overlap — overlap between adjacent chunks
• Ingestion batch size — how many chunks are processed per batch during upload
• Embedding batch size — how many chunks are embedded per batch

Readers

Which extractor is used per file format (for example Markdown, PDF) before content is split for search.