Rust Glancer: A Rust language server built to use 100x less RAM than rust-analyzer

JeremyCMorgan · x · 2026-09-02

A developer has spent 4 months building Rust Glancer, a Rust LSP implementation focused on low memory usage: it targets <100MB for reasonable projects and supports immediate indexing after restart — no re-indexing needed when the editor restarts. In a recorded demo RAM stayed under 100MB, and it runs well on an 8GB MacBook Pro M1 from 2020.

In benchmarks, base indexing is on par with rust-analyzer (6s vs 7s on an M1) while full indexing is faster (9s vs 14s). The author admits it is incomplete with known bugs, but it already has a full indexing pipeline with type inference and a chalk-based trait solver, plus common LSP actions like goto definition, hover, inlay hints and completions, available via a VS Code extension. The design notes on what it deliberately skips to save memory are the most useful part for anyone who has watched an LSP eat 8GB on a large workspace.

Original post →

More from coding & agent

coding & agent channel →