Docs
Everything you need to know about Stackmatch, ranks, and our methodology.
About Stackmatch
Stackmatch helps you find stackmates — developers and organizations who share your unique dependency graph. By scanning package.json files across GitHub, we build stacker-level stack fingerprints and surface meaningful overlaps.
How It Works
Every stacker (user or organization) scanned by Stackmatch goes through a multi-stage aggregation process to identify their core stack.
- Repository Discovery — We identify the top 20 non-fork repositories by star count for a given stacker.
- Manifest Extraction — We recursively scan these repositories for
package.jsonfiles, identifying both root and nested dependencies. - Fingerprint Construction — We aggregate all discovered packages into a single "stack fingerprint" for the stacker, distinguishing between direct dependencies and devDependencies.
- Similarity Engine — To find stackmates, we compare fingerprints using a modified Jaccard similarity index, weighted by package popularity, package signal strength, and repo-level usage.
Stackmatch removes hard-noise boilerplate such as @types/*, @babel/*, eslint config/plugin packages, and prettier config/plugin packages from match scoring. Core tooling such as eslint, prettier, typescript, vitest, and @biomejs/biome stays visible in stack views, but contributes less match evidence than product or framework dependencies.
Data Privacy
Stackmatch uses GitHub OAuth for identity and public repository package analysis.
- Public by default — Public repositories are scanned by default. Standard sign-in does not request private repository access.
- No source storage — We parse package manifests to aggregate dependency names and counts. We do not clone or store source code.
- Profile control — You can hide your profile from discovery with Ghost Mode and clear private aggregate data from your profile controls.
- Optional private sync — Private repository analysis requires installing the Stackmatch GitHub App, where you choose which repositories to grant access to.
- Private aggregate storage — For private analysis, Stackmatch stores aggregate dependency package names/counts and sync status keyed to your GitHub login. It does not store private source code, file paths, commit messages, commit SHAs, or private repository names.
Data Scope & Limitations
| Factor | Stackmatch | GitHub Graph |
|---|---|---|
| What's counted | Packages in package.json | Commits, PRs, Issues |
| Repository Limit | Top 20 public non-forks | All contributed repos |
| Analysis Depth | Full monorepo support (nested manifests) | Usually root-level only |
| Update Frequency | On-demand scan or 24h cache | Real-time |
Stackmatch focuses on intent through dependencies. The packages you choose to build with are the strongest signal of your engineering DNA.
Community
Stackmatch is an MIT-licensed open-source project operated by David Dias Digital. The service is built for developers who want better ways to find stackmates, while the source stays open for review, contribution, and community trust.
If you have ideas for new matching algorithms or want to improve the product, share feedback from your profile or reach out through the contact page.
Open Source Success Loop
Stackmatch is designed to grow like a healthy open-source project, not just a directory.
- Easy first contribution — Improve docs, tests, package pages, match explanations, accessibility, or onboarding copy.
- Visible recognition — Contributors are credited through the project contributor list and public release notes when work ships.
- Public roadmap pressure — Product themes stay visible so developers and companies can see what is being improved next.
- Transparent sponsorship — Sponsors support the developer graph and aggregate ecosystem surfaces. They do not receive private repository access, private inbox access, or hidden lead lists.
- Maintainer-friendly artifacts — Package pages and organization pages should become useful evidence for maintainers, DevRel teams, and developers deciding where to contribute.