We introduce a conditional flow-matching method for missing data imputation, which matches or outperforms a wide range of existing methods on tabular and time-series data.