We exploit different routes for encapsulation of food additives, such as minerals or vitamins, in a polymeric capsule. The added active ingredients should remain inside the capsule for at least a year in an aqueous environment (e.g. a dairy product), since sensory properties or functionality of the ingredients may otherwise be affected. However, after intake the active compound should readily (within 1 h) be released due to the acidic environment in the stomach. First, we propose a phenomenological model in order to study how a polymeric matrix may limit the diffusion of incorporated active molecules. The relation between the release rate of the active compound and its molecular weight is elucidated. Second, the desired capsules may be obtained by specific binding between subunits within the capsule and the active ingredient. We show two examples that rely on this mechanism: amylose-lipid complexes and mixed metal hydroxides. Amylose is able to form inclusion complexes with various types of ligands, including iodine, monoglycerides, fatty acids and alcohols, where the hydrophobic parts of the ligands are entrapped in the hydrophobic helical cavity of amylose. Mixed metal hydroxides are a versatile class of inorganic solids that consist of sheets of metal cations that are octahedrally surrounded by hydroxide molecules. In between these layers anionic species compensate for charge neutrality. In this way, various metal cations (minerals) may be incorporated with a high loading, and negatively charged actives may be placed between the layers. Upon digestion the particles dissolve and the ingredients are digested. Finally, we show that nature has already developed many intriguing capsules.