EMOJICON PUPPETING

A method, comprising receiving, by a device, a plurality of image frames comprising a facial expression of a person. Receiving, by the device, depth information comprising the facial expression of the person. Determining, by the device, the facial expression of the person based on the plurality of image frames and the depth information. Animating, by the device, an avatar based at least in part on the facial expression of the person. Causing a first version of the animated avatar to be rendered, by the device, based on the facial expression. Transmitting recipient information from the device to a message system for a transmission to a recipient computing device associated with the recipient information. Receiving, at the device, a response from the message system prior to delivery of the rendered first version of the animated avatar to the recipient computing device. Causing, following the response from the message system, a second version of the animated avatar to be rendered at the device based on the facial expression, the second version corresponding to one or more capabilities of the recipient computing device associated with the recipient information. Transmitting the rendered second version of the animated avatar from the device to the recipient computing device without transmitting the first version of the animated avatar from the device to the recipient computing device.