ZetaProse

F.A.Q.

Why don't you include spellchecking and grammar correction?

There are already plenty of other free services that do this.


Why don't you have such-and-such a feature?

Take a look at the feature roadmap. Maybe you'll find it listed it there. If not, drop me a line and I'll consider it.


Why is such-and-such a word marked as an adverb? How does the adverb detection work?

Adverb detection uses the Averaged Perceptron part-of-speech tagger. A part-of-speech tagger is an algorithm that adds tags ('noun', 'verb', 'adverb', etc.) to each word in a text. The Averaged Perceptron uses features like the word suffix, prefix, and the surrounding words to predict the correct part-of-speech. You can read more about the Averaged Perceptron tagger here.


How is my private data handled?

No information is stored about your text or its content, and session logs are deleted daily from the server. No one ever sees your text except you.


What's the tech stack behind this site?

The site is a Django webapp hosted on an AWS EC2 instance. The backend is built on several python libraries: nltk for basic NLP (tokenization, stemming, lemmatization, part-of-speech tagger, etc.), spacy, gensim, and matplotlib. The "artificial intelligence" in the Just for Fun tab is an n-gram generative language model with Kneser-Ney smoothing. The sentiment analysis uses a dictionary based on the work of Hu and Liu.