Can an algorithm create typewriter art that captures the tonal range and the spatial resolution of photorealistic images? Typewriter art, and more recently ASCII art, uses textual characters to create images. Two general styles - tone-based and shape-based art - have been long pursued by typists and programmers alike.\newline Taking advantage of the ability to move the paper in the typewriter, we employ multiple, offset layers that can be typed repeatably with no artistic skill. The increased tonal range and shape resolution this offers allows for 'realist', if not quite realistic, photo printing using only a typewriter. We use two metrics which compare the typed output (mockup) against the input photograph (target) for realism: asymmetric mean squared error (AMSE), which prefers tone-matching; and structural similiarity index measure (SSIM), which prefers shape-matching. Blending these metrics generates the most compelling results. Since the typed characters overlap, the choices cannot be treated independently. Greedy search variations and simulated annealing are used to find an approximate optima in this large search space. Due to the discrete nature of the search space, we move in only one dimension at a time, by considering the effect of swapping out a single character for another. Typed results have a high degree of detail and a wide tonal range, while still communicating the image's origins as typewritten text.