Today, Javascript runtimes intend to process data both at server and client levels. In this paper, we study how Rust and Web Assembly can contribute to implement efficient Semantic Web libraries for the Javascript ecosystem. We propose WasmTree, a new implementation of the RDFJS Store and Dataset interfaces in which the processing is carefully split between the Web Assembly and Javascript layers. Experiments show that our best setup outperforms state-of-the-art implementations for fine-grained data access and SPARQL queries.