How can I check token usage per document when indexing with File Search?

Gemini API
,
1

Hi,

I’m using Gemini File Search and trying to understand the token usage (cost) during indexing.

What I want to know specifically is:

  • Token usage per document at indexing time
  • Or token usage per generated chunk
  • Or any usage / billing-related metadata returned by the API

So far, I’ve checked:

  • uploadToFileSearchStore (raw REST API response, including resumable upload)
  • Operations.get response
  • fileSearchStores.documents.get response
  • grounding_metadata during generateContent

But I couldn’t find:

  • Chunk contents created during indexing
  • Token count per chunk
  • Total tokens consumed when uploading a document

From the API docs, it looks like these details are not exposed, but I want to confirm:
:backhand_index_pointing_right: Is there currently any supported way (API, metadata, or header) to get token usage per document for File Search indexing?

If not, is the intended approach to rely only on Cloud Billing / Usage reports?

Thanks in advance.