We present the largest homogenous dataset of grain prices for four major types of grain for pre-industrial Germany covering 70 cities with a total of 259 time series. This article describes the sources and the methodology followed to create calendar-year-based price series in grams of silver per litre. The dataset is organised and documented decentrally by means of one excel file for each city. To aggregate information for all cities in one file, future users can either use spreadsheet software or open-source software code that creates the national-level file. We use the dataset to show how price dispersion developed across more than 300 years. The dataset facilitates future research on market integration, storage, historical demography, and economic impacts of climate change.