Why We Built a Search Runtime in Rust and Compiled It to WebAssembly

Moss is an embedded retrieval system designed to run alongside AI agents in browsers, servers, phones, and edge environments, allowing queries to execute locally without a cloud request for every lookup. Its architecture separates cloud-based document ingestion, index building, and update distribution from local runtime functions such as query embedding, filtering, ranking, and result retrieval, enabling continued search during temporary connectivity loss after an index is loaded. Built in Rust for memory control and safety, with WebAssembly support for browser deployment, Moss shares a common search core across platforms while using platform-specific bindings and builds. Startup performance depends on downloading modules, models, and indexes, while memory and storage needs vary with index content, metadata, models, and update handling; automatic refresh can replace indexes in the background but may require capacity for both versions temporarily. Published benchmarks report low local query latency on a specified MacBook Pro workload, but do not establish browser performance, startup costs, or resource use across all targets. The approach can reduce per-query network dependence, though browser deployment introduces download, security, and integration constraints, while native deployment requires operating-system and architecture compatibility; applications must choose an index size and runtime location that fit their available resources.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.