Maki, the efficient coder Maki, a Rust-based coding agent, parses 15 languages into structural skeletons via tree-sitter, costing 59 tokens per turn while saving 224 tokens on reads, according to the project's documentation. Maki also runs a sandboxed Python interpreter where tools are async functions, and hides Datadog's MCP server's more than 100 tool definitions behind a single search tool that loads only what the model requests. The agent ships as a native binary with a 60 FPS Rust TUI, a Lua plugin API, and tree-sitter-parsed bash permissions that flag both git and rm in chained commands. Where tokens go index index Parses 15 languages into skeletons: imports, type defs, function signatures with their line ranges. Costs 59 tok/turn, saves 224 on reads. Reads were ~65% of my tokens, so this one is big. code execution exec A sandboxed Python interpreter where every tool is an async function. The model gathers 50 reads, greps them, prints the 3 lines that matter. The rest never touches your context. tool search https://maki.sh/docs/mcp/ tool-search Datadog's MCP server has over 100 tools. Every definition sits in your context on every request, used or not. Maki hides them behind one search tool and loads what the model asks for. task The model picks weak, medium, or strong for each subagent. Haiku-tier for grep-heavy research, opus-tier for architecture. You get a summary, not the transcript. compaction Long sessions get compacted: images and thinking blocks go first, then old turns get summarized. The system prompt and tool descriptions are short too. index: read less, know more Instead of reading full files, index parses with tree-sitter and returns a compact skeleton. The model sees the structure, then reads only the lines it needs. main.rs hover to restore use std::fs;use clap::Parser;use color eyre::Result; derive Parser struct Args { paths: Vec