Skip to main content
Back to Help

Syncing vs indexing

When Nomic connects to an external system, files move through two important stages: syncing and indexing. A file can be visible after syncing but not fully searchable until indexing finishes.

Syncing

Syncing brings file information into Nomic from a connected system.

Syncing usually includes:

  • Folder names
  • File names
  • Metadata
  • Permissions
  • Version information

After sync, the file appears in Nomic, but its contents are not searchable until it is indexed.

Indexing

Indexing makes file contents searchable. Nomic extracts text, tables, drawings, images, and other document structure so Assistant and workflows can use the file.

Syncing does not index. Run Index Folder from the action menu on a synced folder — that one request keeps covering files the sync brings in later. (Egnyte syncs can be configured by an admin to index automatically.) Individual files use Index Document.

Assistant can start working while files are still indexing, especially when files are attached directly to the session. Search quality, citations, and large workflow reviews are strongest after the relevant files show Indexed.

Indexing usually happens once per file version. If a new version is uploaded or synced, that version may need a new indexing pass.

Why this distinction matters

If a file appears in Nomic but Assistant does not use it, check:

  • Is the file still indexing?
  • Did indexing fail?
  • Is the file included in the current project, folder, workflow, or session?
  • Does the user have access to the file in the source system?
  • Is the workflow using an old file version?

Cost and reuse

Indexing has a cost because Nomic processes the file. After a file is indexed, future Assistant questions and workflows can reuse that indexed content.

To avoid unnecessary spend:

  • Reuse files already in the project or drive.
  • Avoid uploading duplicate copies of the same large document.
  • Use new versions only when the content has actually changed.
  • Start workflows from indexed files when possible instead of repeatedly attaching large unindexed files.