Skip to main content

​
What is Brain Access?

Brain Access controls whether the AI assistant can search and cite your vault documents. When Brain Access is enabled for a vault or file, the Brain can find and reference that content when answering questions in chat.

​
Per-Vault Control

​
Toggling Brain Access on a Vault

1

Find the vault in the sidebar

In the Vault left sidebar, find the vault you want to configure.
2

Open the vault dropdown

Click the three-dot menu that appears on hover.
3

Toggle Brain: ON/OFF

Click the Brain: ON/OFF toggle. A green switch means Brain has access; gray means it doesn’t.
When Brain Access is OFF for a vault:
  • A “NO AI” badge appears next to the vault name in the sidebar
  • None of the files in that vault are searchable by the Brain
  • Files are still stored and accessible to users — they just aren’t included in AI search
Alex’s screenshot note: Take a screenshot of the vault sidebar showing one vault with the “NO AI” badge next to its name, and the dropdown menu open on another vault showing the Brain toggle in the ON state.

​
Per-File Control

You can also control Brain Access on individual files:

​
Right-Click Method

  1. Right-click a file
  2. Select Add to Brain (to enable) or Remove from Brain (to disable)

​
Status Indicator

In list view, the Status column shows a brain icon for files that are indexed and Brain-accessible.

​
When to Disable Brain Access

Common reasons to disable Brain Access:
  • Sensitive documents — HR files, legal documents, salary data that shouldn’t appear in AI responses
  • Drafts — Work-in-progress documents that aren’t ready to be cited
  • Archived content — Old files that might confuse the AI with outdated information
  • Personal files — Files you don’t want surfaced in team conversations
A good practice is to keep Brain Access ON for all finalized, team-relevant documents and OFF for sensitive or draft content.

​
How Indexing Works

When Brain Access is enabled, files go through this pipeline:
1

Text extraction

The file’s content is extracted (OCR for images, text parsing for documents).
2

Embedding generation

The text is converted into vector embeddings using OpenAI’s text-embedding-3-large model.
3

Pinecone storage

Embeddings are stored in Pinecone, organized by vault namespace.
4

Searchable

The Brain can now find this document through semantic search when answering questions.
This process happens automatically when you upload a file to a Brain-accessible vault.