[eng] Plasmonic enzyme-linked immunosorbent assays (ELISA) using the localized surface plasmon resonance (LSPR) of metal nanoparticles has emerged as an appealing alternative to conventional ELISA counterparts for ultrasensitive naked-eye detection of biomolecules and small contaminants. However, batchwise plasmonic ELISA involving end-point detection lacks ruggedness inasmuch as the generation or etching of NP is greatly dependent on every experimental parameter of the analytical workflow. To tackle the above shortcomings, this paper reports on an automatic flow methodology as a reliable detection scheme of hydrogen peroxide related enzymatic bioassays for ultrasensitive detection of small molecules. Here, a competitive ELISA is combined with the in-line generation of plasmonic gold nanoparticles (AuNPs) followed by the real-time monitoring of the NP nucleation and growth rates and size distribution using a USB miniaturized photometer. Glucose oxidase was labeled to the secondary antibody and yielded hydrogen peroxide that acted as the measurand and the reducing agent of the Au(III)/citrate system in the flow network. High-throughput plasmonic assays were feasible by assembling a hybrid flow system composed of two microsyringe pumps, a perfluoroalkoxy alkane reaction coil, and a 26-port multiposition valve and operated under computer-controllable flow conditions. The ultratrace determination of diclofenac in high matrix samples, e.g., seawater, without any prior sample treatment was selected as a proof-of-concept application of the flow-based platform for determination of emerging contaminants via plasmonic ELISA. The detection limit (0.001 μg L-1) was 1 order of magnitude lower than that endorsed by the first EU Watch List for diclofenac as a potentially emerging contaminant in seawater and also than that of a conventional colorimetric ELISA, which in turn is inappropriate for determination of diclofenac in seawater at the levels endorsed by the EU regulation. The proposed automatic fluidic approach is characterized by the reproducible timing in AuNPs nucleation and growth along with the unsupervised LSPR absorbance detection of AuNPs with a dynamic range for diclofenac spanning from 0.01 to 10 μg L-1. Repeatability and intermediate precision (given as normalized signal readouts) in seawater were <4% and <14%, respectively, as compared to RSDs as high as 30% as obtained with the batchwise plasmonic ELISA counterpart.