Performance for large number of items

[hhimanshu]

Hello,

I just stumbled upon this library. I wanted to know if there are any recommendations on the items size when using this library?
I have over 3 million items with following structure

[
  {
    "food": "Food1",
    "category": [
      "Herbs"
    ],
    "u": "/xyz",
    "synonyms": [
      {
        "l": "X",
        "n": "Y"
      }
    ]
  }
]

My questions are

1. Is it feasible to use flexsearch on browser with these many items?
2. Can we pre-index in a file and load on the client using an API? The index is immutable and wondering if I could save resources on the client (browser)

Please share your thoughts. Thank you