Whole-cell biosensors are powerful diagnostic tools for detection of various analytes, including proteins which have difficulty in transporting across the cell membrane. Here, we developed a novel biosensor based on the agglutination reaction by using an E. coli chassis displaying nanobodies (VHHs) as the detection element for selectively binding to a target of interest. As a proof-of-concept, we demonstrated that this design was able to detect a model protein having two antigenic determinants in a semi-quantitative manner with a detectable change in the output, which was easily visualised using the naked eye. Subsequently, the designer cells were re-engineered to develop tests for clinically relevant biomarkers by the display of specific VHHs against human fibrinogen (Fib) and 3-phenoxybenzoic acid (3-PBA). Fib is a bivalent protein so detection was achieved in the same manner as the prototype by observing cell agglutination. 3-PBA is a small molecule, therefore, detection was achieved by modifying the platform based on a competitive ELISA technique. Binding with a 3-PBA-protein conjugate, allowed cell crosslinking, which was disrupted by the presence of free 3-PBA, thus resulting in cell pellets. The optimised cell biosensors detected both analytes at nanomolar concentrations and exhibited robust function in complex sample backgrounds such as human plasma and synthetic urine. Furthermore, we studied the feasibility of using the bacterial LuxIR quorum sensing as a quantitative output design for differentiating cell agglutination versus pellets by monitoring the production of the fluorescent reporter. Despite the verified functional constructs, there was no signal difference between the two states, possibly due to a similar spatial cell distribution in a 96-well plate. This agglutination-dependent E. coli biosensor was shown to be a promising platform for identifying both protein and chemical analytes. Therefore this simple, affordable design could be developed as a potential diagnostic tool for field use.