Accurate
and rapid blood typing plays a vital role in a variety
of biomedical and forensic scenarios, but recognizing weak agglutination
remains challenging. Herein, we demonstrated a flipping identification
with a prompt error-discrimination (FLIPPED) platform for automatic
blood group readouts. Bromocresol green dye was exploited as a characteristic
chromatography indicator for the differentiation of plasma from whole
blood by presenting a teal color against a brown color. After integrating
these color changes into a quick-response (QR) code, prompt typing
of ABO and Rhesus groups was automatically achieved and data could
be uploaded wirelessly within 30 s using a commercially available
smartphone to facilitate blood cross-matching. We further designed
a color correction model and algorithm to remove potential errors
from scanning angles and ambient light intensities, by which weak
agglutination could be accurately recognized. With comparable accuracy
and repeatability to classical column assay in grouping 450 blood
samples, the proposed approach further demonstrates to be a versatile
sample-to-result platform for clinical diagnostics, food safety, and
environmental monitoring.