The following demo builds the inverted index on my personal tweet archive (1916 tweets through August 5, 2015) using IndexedDB. Everything is designed to work on Chrome, so YMMV on other browsers.

If you pop open your JavaScript console, you'll find the variable tweets containing the tweets that we're going to index:

> tweets
Array[1916]

These are JSON objects generated by Twitter's "Download your tweet archive" service, so you can manipulate the collection as normal JSON objects:

> tweets[1912].text
"Moving day! http://t.co/yBse0eJmnq"

> tweets[1912].created_at
"2015-08-05 12:19:21 +0000"

Okay, now we're ready to index. First click the "Build Index" button, then click the "Build DF Table" button. You can watch the indexing progress in the JavaScript console. After that, click the "Get Index Stats" button: there should be 27308 postings and 6413 terms.

Build Index Build DF Table Get Index Stats Delete index

Current Index Status

Number of Postings
Number of Terms (in DF table)