jahala/tilth
View on GitHub ↗ ★ 357 Rust
Smart code reading for humans and AI agents. Definitions, callers, dependencies and function-level diffs from tree-sitter. One binary, no index. MCP server + CLI. Install: cargo install tilth or npx tilth
15 mentions on Hacker News by 3 people: 5 stories, 10 comments · first 2026-02-09 · latest 2026-09-12
20082026
-
I've gotten good usage out of a structured search tool (not mine, by someone else here) called Tilth: https://github.com/jahala/tilth It combines search with tree sitter grammars so the results can annotate usage vs definitions, cite line…
comment by hedgehog · 2026-09-12 · in RTK reports token savings, but our cost benchmarks disagree -
Its on the way! https://github.com/jahala/tilth/pull/151
comment by jahala · 2026-06-21 · in The Token Compression Illusion: Why I'm Skeptical of RTK -
Would sincerely love to hear your thoughts on https://www.github.com/jahala/tilth - it’s a different approach than RTK, benchmarked to reduce cost per correct answer by ~40%
comment by jahala · 2026-06-18 · in The Token Compression Illusion: Why I'm Skeptical of RTK -
…competition, but would be happy to put the numbers up if any token whales feel like having a go. https://github.com/jahala/tilth/tree/main/benchmark
comment by jahala · 2026-06-06 · in Show HN: Lowfat – pluggable CLI filter that saved 91.8% of my LLM tokens -
Related: https://github.com/jahala/tilth Edit: recent HN discussion: https://news.ycombinator.com/item?id=46952321
comment by te0006 · 2026-05-19 · in Show HN: Semble – Code search for agents that uses 98% fewer tokens than grep -
…range for function definitions, provide file outlines with line numbers on a file name match, etc. https://github.com/jahala/tilth
comment by hedgehog · 2026-04-27 · in Show HN: OSS Agent I built topped the TerminalBench on Gemini-3-flash-preview -
I built tilth ( https://github.com/jahala/tilth ) much for this reason. Couldn't bother with RAG, but the agents kept using too many tokens - and too many turns - for finding what it needed. So I combined ripgrep and tree-sitter and some…
comment by jahala · 2026-04-04 · in We replaced RAG with a virtual filesystem for our AI documentation assistant -
…code reading for humans and AI agents. Reduces LLM token use and cost by ~40% (benchmarked) https://github.com/jahala/tilth - mrkd: Native macOS .md viewer (+preview in Finder) that imports iTerm2 and VSCode themes…
comment by jahala · 2026-03-09 · in Ask HN: What Are You Working On? (March 2026) -
▲ 4
Show HN: Tilth v0.5.0 –> ~40% cheaper AI code navigation (160 runs, 3 models)story by jahala · 2026-03-08 · 2 comments
-
…much for the same reasons. Its MIT licence, open source and saves you 30% token use (benchmarked) https://github.com/jahala/tilth
comment by jahala · 2026-03-06 · in Claude Code told me what tools it needs to work faster -
▲ 2
Show HN: Reduce LLM token use by ~30% with this MCP/CLI tool(Claude benchmarked)story by jahala · 2026-03-05 · 1 comment
-
▲ 2
Show HN: Tilth v0.4.1 – 29% cheaper Sonnet, 22% on Opus (benchmark: 114 runs)story by jahala · 2026-02-17 · 0 comments
-
▲ 3
Show HN: Tilth v0.3 – 17% cheaper AI code navigation (279 runs, 3 Claude models)story by jahala · 2026-02-14 · 0 comments
-
I implemented this hash (read and edit) approach in tilth if you want to test it out. https://github.com/jahala/tilth its on npm and cargo: - cargo install tilth - npx tilth then tilth install claude-code/windsurf/cursor --edit (--edit…
comment by jahala · 2026-02-12 · in Improving 15 LLMs at Coding in One Afternoon. Only the Harness Changed -
▲ 10
Show HN: Tilth – I spent tokens so my agents would stop wasting them (~4k Rust)story by jahala · 2026-02-09 · 2 comments