Knowledge
Knowledge spaces and content memory
Organize sources, documents, and searchable context instead of accumulating bookmarks.
A knowledge space groups material around a durable topic, client, or content program. It also constrains retrieval scope and creation permission.
Space design
- Name spaces after stable subjects, not temporary formats.
- Treat source evidence and editable documents as distinct objects.
- Use separate spaces or workspaces for client and permission boundaries.
- Regularly review failed imports, duplicates, and text without provenance.
What GBrain does
The backend uses GBrain for pages, chunks, vectors, facts, and task queues. Product entry points never connect directly to its database; Backend brain infrastructure owns retrieval and writes.
Users should decide knowledge scope and citation quality, not tune vector index internals.