The covid-19 outbreak left many countries have no choice but turn to online education. Turkish students, who were among those who were affected, faced difficulties in improving their English as the opportunity to have face-to-face feedback was not available. In this work, we build and open-source a dataset for grammatical error correction composed of essays written by Turkish students from different universities capturing the errors Turkish native speakers tend to make. We utilize the dataset and build a model, which we deploy along with a web interface.