A probabilist spelling corrector that runs in the browser that let to train, correct text and diff results right in the browser.
$ git clone https://github.com/loretoparisi/spelling-corrector.git
$ cd spelling-corrector/demo
$ python -m SimpleHTTPServer 8181
and point your browser to https://localhost:8181
The spelling corrector page looks like
Click on train first to train the spell checker. The console output will tell you the the stopwords loaded and the training time. To run a test of the trained model click on Test 1 or Test 2. You should see an output like
{"bad":130,"n":400,"pct":"68","unknown":43,"secs":"6"}
To check the spelling on a single word, type the word in the Single Term box and click on correct. To check the spelling of the whole text, type the text in the left text box below (or leave the example's one) and click on full text correct. In the left pane you have the input source text - possibly mispelled, in the middle box you have the corrected text, and in the right box you can see the differences.
This project is based on speller by past, that implements the insane famous Peter Norvig's spell-checker. For text comparison it uses JSDiff