Embodiments are directed to providing an image in response to a request for an image and to requesting an image for a specified context. In one scenario, a computer system stores multiple image representations of a corresponding visual asset identified by an identifier. The image representations include a vectorized representation of the visual asset and/or a rasterized representation of the visual asset. The computer system receives a request that includes an identifier for one of the visual assets and usage parameters. The computer system evaluates at least one of the usage parameters to identify an image representation of the identified visual asset that substantially corresponds with the usage parameters. The computer system then obtains the identified image representation of the corresponding visual asset based on the evaluation and provides the obtained image representation in response to the request.