Tweets often contain a large proportion of abbreviations, alternative spellings, novel words and other non-canonical language. These features are problematic for standard language analysis tools and it can be desirable to convert them to canonical form. We propose a novel text normalization model based on learning edit operations from labeled data while incorporating features induced from unlabeled data via character-level neural text embeddings. The text embeddings are generated using an Simple Recurrent Network. We find that enriching the feature set with text embeddings substantially lowers word error rates on an English tweet normalization dataset. Our model improves on stateof-the-art with little training data and without any lexical resources.