buckaroo.pm - Find C++ packages

Finding a C++ library has always been harder than it should be. C++ build tooling is chaotic and unstandardised — it is common for projects to reinvent the wheel or vendor and modify dependencies just to make them compile. There is no npm, no PyPI. buckaroo.pm is a companion website for the Buckaroo package manager that tries to change this: a searchable index of around 400 packages available through Buckaroo, all sourced from GitHub by crawling buckaroo.toml manifests. Most of the packages in the index were created and maintained by LoopPerfect up until around 2021.

What buckaroo.pm Does

At its core, buckaroo.pm is a package explorer. It indexes packages found on GitHub — specifically any repository containing a buckaroo.toml manifest — and presents them in a way you can actually navigate:

• Search any package by name, topic, or keyword
• Browse a package's metadata: description, available versions, dependencies
• Explore the dependency graph — see what a package depends on and what depends on it

The dependency graph view is particularly useful in the C++ world, where understanding transitive dependencies before you commit to a library can save hours of integration pain.

Instant Search with Precompiled Fuzzy Search

The most visible feature is the search box: type anything and results appear immediately, with no network round-trip.

This is possible because the search index is precompiled at build time and shipped as a static asset. Rather than querying a server on each keystroke, the entire index is downloaded once when the page loads. Fuzzy matching then runs entirely in the browser — in JavaScript, against the in-memory index — so results are instant regardless of network conditions.

The index encodes package names and descriptions in a structure optimised for fuzzy lookup: small edits (a missing letter, a transposition) still surface the right package. This is the same class of problem explored in the Fuzzy Search article — edit distance, trie-based pruning, and tolerance thresholds that balance recall against precision.

Because the index is static, the entire site can be served from a CDN with no backend. There is no database, no search API, no infrastructure to maintain. The cost of hosting is effectively zero, and the search latency is bounded only by the user's machine — not a server on the other side of the Atlantic.

Dependency Graph Exploration

Each package page includes an interactive dependency graph showing the immediate and transitive dependencies of that package. This turns what would normally be a cat BUCK.lock | grep ... exercise into a visual exploration.

The graph is rendered client-side from the manifest data embedded in the static page. Click a node to navigate to that package, see its own dependency graph, and trace paths through the ecosystem. It is the kind of tool that makes the shape of C++ dependency management legible in a way that a command-line tool alone cannot.

Why This Matters

The fragmentation of the C++ ecosystem is not just a tooling problem — it is a discoverability problem. C++ projects often reinvent or vendor the same libraries because finding a reusable, compatible version is too painful. buckaroo.pm is an attempt to give the C++ world the kind of package browsing experience that developers in other ecosystems take for granted: instant search, browsable metadata, and a clear picture of what you're taking on before you add a dependency. The index is small — around 400 packages, mostly maintained by LoopPerfect until 2021 — which itself reflects how difficult adoption proved to be. But as a proof of concept for what a discoverable C++ ecosystem could look like, it did the job.

Related

• Buckaroo — The package manager that buckaroo.pm indexes.
• Buckaroo: Conflict-Driven Dependency Resolution for C++ — How Buckaroo resolves the dependency graph once you've chosen your packages.
• Fuzzy Search: Typos, Tries, and the Algorithms Behind Instant Results — The algorithms powering the instant search experience.