stout vs Homebrew: 10-100× faster package installs
stout is a Rust, Homebrew-compatible package manager. A pre-computed SQLite index with FTS5 search, partial fetches and parallel bottle downloads.
stout is a Rust-backed drop-in for
Homebrew. Same API, faster hot paths.
Install: see docs.
10-100× faster than Homebrew on common operations.
stout vs Homebrew: the facts
10-100× faster than Homebrew on common operations.
Bounded by Ruby startup + linear scans.
Pre-computed SQLite index with FTS5 full-text search.
Live git index; `brew search` is grep on the formulae repo.
Fetches only what it needs.
Updates the full index on every `brew update`.
Parallel bottle downloads.
Sequential by default.
Compatible with Homebrew formulae.
Native Homebrew.
No Ruby runtime.
Requires Ruby (slow to start).
MIT licensed.
BSD-licensed.
Benchmarks
Reproduction instructions in the project README. Numbers measured on AMD Ryzen 9 7950X, 64GB DDR5, NVMe SSD, Python 3.12.
| Metric | stout | Homebrew |
|---|---|---|
| `brew update` (cold) | ~1-2s | ~10-30s |
| `brew search <term>` | ~50ms | ~1-3s |
| `brew install <pkg>` (already in index) | ~2-5s | ~5-20s |
Why an index beats a git checkout
Homebrew answers questions about packages by consulting a git working copy of the formulae repository and evaluating Ruby. That design is flexible — a formula is a program — but it means that searching is closer to a scan than a lookup, and that updating means reconciling a repository rather than fetching a diff of what changed.
stout pre-computes the same information into a SQLite database with an FTS5 full-text index, so a search is an index lookup and an update fetches only what moved. Removing Ruby from the path also removes interpreter startup from every single command, which is why the gap is largest on the short operations you run most often rather than on a long compile.
Compatibility, and what that word covers here
Compatibility means stout consumes Homebrew formulae and fetches from the same bottle mirrors, so the artefact you install is the artefact Homebrew would have installed. It does not mean every Homebrew subcommand, tap arrangement and edge-case behaviour is reproduced. Treat the formula and bottle compatibility as the contract, and check anything beyond that against the current release.
The practical consequence is that stout is easy to trial: install it alongside Homebrew, use it for search and update where the speed difference is most visible, and keep brew for anything it does not cover yet. Nothing has to be migrated for that to work.
Where Homebrew is still the safer choice
Homebrew has a decade of accumulated taps, a very large maintainer community, and a support answer for essentially every question a user can ask. If you depend on a tap or workflow stout does not yet handle, or if you are setting up machines for people who will need to search the web for help, that ecosystem advantage is real and is worth more than the time saved on brew update.
The other case is a machine where package operations are rare. If you install a toolchain once a quarter, a faster package manager is solving a problem you do not have. See the rest of the developer tooling family for loops where the repetition is what makes the speed matter.
When to use stout
- You use Homebrew and are tired of waiting on `brew update` and `brew search`.
- You want a permissively-licensed package manager (MIT vs BSD) that is Homebrew-compatible.
- You script machine setup and run package operations often enough that per-command latency adds up.
When NOT to use stout
- You depend on a Homebrew tap that stout doesn't support yet.
- You install a toolchain once a quarter — a faster package manager solves a problem you do not have.
- You need the breadth of Homebrew community answers more than you need speed.
Frequently asked questions
Will stout run my existing Homebrew formulae?
Yes — stout is Homebrew-compatible. Existing formulae install with `stout install <formula>`.
Does stout use the same bottle binaries as Homebrew?
Yes — stout fetches from the same bottle mirrors that Homebrew uses.
Can I run stout and Homebrew side by side?
Yes, and that is the recommended way to trial it. Use stout for search and update where the difference is most visible and keep brew available for anything stout does not cover.
Does stout compile from source?
It prefers the same prebuilt bottles Homebrew uses. Where no bottle exists the formula still describes a source build, and that path is bounded by the compiler, not by the package manager.