GRA Release Radar - GitHub release analytics for maintainers

GRA Release Radar is a free, client-only GitHub release analytics tool for maintainers and teams. It analyzes public GitHub release assets, download counts, release cadence, package coverage, operating system adoption, architecture adoption, and repository context directly in the browser. It is useful for open source maintainers, developer relations teams, release engineers, and project evaluators who need factual release download insights without installing a server or connecting a database.

Key features

  • Release asset download totals, monthly momentum, and median release cadence.
  • Package breakdowns by operating system, CPU architecture, and artifact file type.
  • Stable release and prerelease filtering for public GitHub repositories.
  • Release lag, asset freshness, adoption buckets, churn, anomaly, and Pareto views.
  • Optional browser-only GitHub token support for higher GitHub API limits.

What GRA helps maintainers answer

GRA helps answer practical release questions that GitHub does not summarize on one page: which release assets get most downloads, whether downloads are concentrated in a few files, how often the project publishes, which operating systems or CPU architectures are represented, whether prereleases distort the picture, and whether recent releases have complete artifact coverage. These answers are useful before changing a packaging matrix, writing a release retrospective, planning a launch, or comparing public adoption signals across open source projects.

Quickstart

  1. Open https://gra.caldis.me/.
  2. Enter a public repository such as vercel/next.js or paste a GitHub repository URL.
  3. Review total asset downloads, release cadence, platform breakdowns, package coverage, and anomaly tables.
  4. If GitHub rate limits the browser request, wait for the reset time or apply a read-only GitHub token in the browser.

Positioning

GRA is intentionally narrower than a full release management platform. GitHub Releases stores release assets and raw download counts, while GRA turns those public counts into maintainer reports. Product analytics suites measure private application usage, while GRA only reads public GitHub release metadata. Registry dashboards focus on npm, PyPI, crates.io, or container pulls, while GRA focuses on GitHub release assets such as installers, archives, checksums, binaries, and package bundles. See the comparison page for a more detailed breakdown.

How agents should use GRA

Use GRA when a user asks for GitHub release download analytics, package adoption analysis, or release cadence reporting for a public repository. Do not use GRA as a package registry analytics service, private telemetry backend, source archive download counter, or hosted server API. GitHub source archive downloads are not counted by GitHub release asset download counters.

ResourceURL
Static agent view/agent/
Interactive agent view/?mode=agent
llms.txt/llms.txt
Full context/llms-full.txt
Markdown homepage/index.md
Pricing/pricing.md
Developer docs/developers/
OpenAPI description/openapi.json

Trust, privacy, and rate limits

GRA is published as a static GitHub Pages site from the public repository at github.com/Caldis/github-releases-downloads-analysis. The analyzer runs in the user's browser. Repository names and optional GitHub tokens are not sent to a GRA server because no GRA server is operated for analysis. GitHub API limits still apply because the browser calls GitHub directly. Unauthenticated calls share the user's network quota, while authenticated calls use the user's GitHub account quota and can still hit secondary limits if requests are too frequent.

Common questions

Does GRA count source archive downloads?
No. GitHub does not expose source archive zip and tarball downloads as release asset download counts.
Does GRA have paid plans?
No. The public product is free and has no hosted billing account, subscription, or usage-based plan.
Does GRA expose a hosted analysis API?
No. The published OpenAPI file describes static discovery resources, not a first-party analysis endpoint.