Abstract Background Lateral flow (LF) tests have employed nitrocellulose membrane since their introduction. Refinement of nitrocellulose and LF production techniques have led to substantial improvements in reproducibility, sensitivity, clarity, and speed of results. The recent surge in demand for LF tests during the COVID-19 pandemic has drawn increased attention to the consequences of using non-biodegradable or non-recyclable materials in test production, requiring disposal via landfill. The nitrocellulose substrate is a critical component for substitution with an equivalent or superior material to address environmental concerns. Additionally, nitrocellulose and the specific molecular recognition reagents, typically antibodies, contribute significantly to the test cost, and whilst LF tests are less expensive than most other formats, challenges remain with respect to providing affordable tests with adequate performance globally in low resource environments. Despite attempts to substitute cellulose for nitrocellulose, principally for reducing cost and more latterly, environmental impact, cellulose is not widely used due to inferior performance, especially in relation to the clarity and quality of the signal line and reagent requirements. Here, we present preliminary data supporting the use of cellulose membrane (paper) as an alternative substrate to nitrocellulose for LF assays. We demonstrate that cellulose-binding module (CBM)-streptavidin fusion proteins for the immobilisation of capture antibodies on cellulose addresses the performance limitations, yielding sensitivity, clarity and speed of result matching current clinical and industry requirements for a qualitative SARS-CoV-2 antigen LF test. Methods We expressed and purified CBM-streptavidin fusion proteins from E.coli to a purity of >90%. Binding to biotinylated proteins and cellulose material was confirmed. An inhouse SARS-CoV-2 antigen assay was adapted for use with the CBM-streptavidin fusion protein on cellulose material. Using methods compatible with current LF test production, we conjugated and dried down 40nm gold (BBI) with anti-nucleocapsid antibodies (Hytest). The biotinylated anti-nucleocapsid capture antibody was dried down on a separate pad. CBM-streptavidin fusion protein was plotted on Whatman 43 paper at the test line. We evaluated assay performance using SARS-CoV-2 nucleocapsid protein and UV-inactivated SARS-CoV-2 virus (NIBSC) in artificial mucus. We further evaluated accelerated stability of the CBM-paper system at 45°C. Results Plotting of the CBM-fusion protein on the test line (0.3 mg/mL, 1 uL/cm) resulted in a tightly focused line, while the control line antibody, without CBM-fusion protein (1 mg/mL, 1 uL/cm), diffused from its deposition line. Additionally, we were able to substantially reduce the amount of capture antibody needed using the CBM-fusion protein system. The cellulose-based assay exhibited a runtime of 30 minutes and an analytical sensitivity of 10^3 pfu/mL. No significant changes in performance were observed after 9 weeks of incubation at 45°C, representing an approximate accelerated shelf-life of 12 months. Conclusion Our results indicate that paper-based tests match sensitivity, clarity, and speed of current qualitative antigen tests on nitrocellulose, with a shelf-life of at least one year. This preliminary investigation demonstrates the viability of using paper as a replacement for nitrocellulose when coupled with CBM-fusion proteins for antibody immobilisation, offering significantly reduced material cost and environmental impact.