Price a media call before making it.

Request body: the body you would send to the generation endpoint, plus kind — image (POST /images/generate/: model, number_of_images, model parameters), video (POST /video/generation/: model, parameters) or music (POST /audio/music/: model, duration_seconds). prompt is not needed.

Returns: {object: "estimate", kind, model, credits, exact, bound, lines: [{item, credits, exact, bound, basis, note?}]}. bound says where the charge lies against the quote: exact, minimum (token-priced images), maximum (a sound effect whose length the engine picks) or approximate (the member's own credit may apply at another rate). note explains it.