ast-grep rewrites tree-sitter's C core in Rust and measures 29.7 percent higher parse throughput

Sources: ast-grep blog, fork repository, HN 49060509
Summary: A post on the ast-grep blog reports a Rust rewrite of tree-sitter's C core. Measured against the unmodified C build normalised to 100, the post reports raw parsing throughput at 129.74, with the peak RSS upper bound rising from 21.41 MiB to 25.70 MiB while the lower bound stayed near flat at 8.48 MiB against 8.42 MiB, tree traversal throughput at 110.16 with RSS up 8.9 percent, and the full ast-grep outline workload at 0.960s user CPU against 1.233s with RSS up 29.8 percent. The post states ast-grep produced an identical outline. The post is part 1 of 4 and states that AI wrote the code. The work lives in a personal fork, HerringtonDarkholme/tree-sitter, and is not upstream tree-sitter.
Why it matters: tree-sitter sits under editors, linters, and code-search tooling, so a memory-for-speed tradeoff in its core is a number every downstream consumer should see before adopting the fork.
Follow-up: Watch for the remaining three parts, for independent benchmarks, and for any upstream tree-sitter response.