DNA nanotechnology has evolved over the past three decades into a promising field, providing programmable molecular building blocks for various applications in biotechnology and synthetic biology. We developed a modular “plug and play” table to explore and optimise various neural network components and evaluated their performance through simulations. We employ chemical reaction network models to simulate these components, revealing insights into their compatibility and performance metrics. Our findings highlight the impact of architectural choices on network functionality and suggest pathways for future advancements. Ultimately, this research emphasises the potential of scalable, in-vivo biological neural networks capable of online learning, opening new avenues for applications in biotechnology, molecular diagnostics, and synthetic biology.