We have created and released a new corpus for german sentiment analysis based on Twitter data. The new corpus and the German word embeddings (plain and optimized) are publicly available.