SiteVerdict — GitHub Repo Analytics & Star History

Is your open-source project winning? Compare it with competitors on the numbers that matter: momentum, maintainer speed and community, not just stars. One row per repository from the official GitHub API: stars and a sampled star history, forks, open issues vs open PRs, PR merge speed, issue close time, contributors, releases and commit activity. Run it on a schedule to track growth between runs and get Slack/webhook alerts for star spikes and new releases.

Who it is for

• Open-source maintainers and DevRel: weekly benchmark against alternatives.
• Investors and analysts: momentum and health signals for many repos or whole orgs.
• Developer-tool marketers: see when a competitor ships a release or goes viral.
• AI agents: "compare these repos" with real, current numbers (see MCP-TOOL.md).

What you get per repository

Field	Meaning
stars, forks, watchers	current totals
starsGained30d, starsGained90d, starHistory[]	sampled from the stargazers API (marked estimated for >40k stars)
starsSinceLastRun, stars7dTracked, stars30dTracked	exact deltas from your own run history
openIssues, openPrs	open issues and PRs split (GitHub's own counter mixes them)
prsOpened30d, prsMerged30d, prsClosedUnmerged30d, medianHoursToMerge	PR velocity (bot PRs excluded)
issuesOpened30d, issuesClosed30d, medianHoursToCloseIssue	issue responsiveness
contributors	contributor count (usernames only if you enable includeContributors)
latestRelease, daysSinceLastRelease, releases90d	release cadence
commits12w, commits52w, daysSinceLastPush, archived	activity and risk
alerts[]	sample-org/fast-router gained 400 stars since the last run (18020 → 18420).

Org scan rows (type: "org-repo") list every public repo with stars, forks, language and last push, and the top N get the full report. Exports: REPORT.html, repos.csv, SUMMARY.

Use cases

1. Competitor benchmark: {"repos": ["you/your-lib", "rival/lib-a", "rival/lib-b"]} weekly, HTML report to the team.
2. Org health review: {"orgs": ["your-org"], "deepAnalyseTopN": 10} to find stale or overloaded repos.
3. Launch radar: daily run over 30 competitor repos with alertOnRelease and Slack.

Token and limits

• No token: 60 GitHub API requests/hour per IP, so the default budget is 55 calls (about 4–5 full reports per run).
• With a token (fine-grained, public repositories read-only is enough): 5,000 requests/hour; default budget 1,000 calls.
• A full report costs about 9–14 calls (6 for the star curve, configurable). The Actor reads GitHub's rate-limit headers and stops cleanly before the limit, reporting what it skipped.

Pricing (pay per event)

Event	Price	When
repo-report	$0.01	one full repository report
org-repo	$0.002	one org scan row
repo-alert	$0.01	one alert
ai-brief	$0.03	AI comparison with your own key (only on success)

Examples: 10 repos weekly = $0.10/week; an org with 300 repos + top 10 deep = $0.70. Not-found repos and rate-limit stops are free. Sample mode is free.

FAQ

Do I need a GitHub token? No for most metrics. Star history is the exception: since October 2026 GitHub answers the stargazers list with HTTP 401 unless you are signed in, so without a token the star-history fields are empty and starHistoryNote explains why. A free read-only token fixes it and raises the limit from 60 to 5,000 requests/hour.

How accurate is star history? The stargazers API returns stars in order with timestamps; we sample a few pages and interpolate, which is accurate for trends. GitHub only serves the first 40,000 stargazers, so curves for bigger repos are marked as estimates. Your own run history (starsSinceLastRun) is exact.

Private repos? Not supported: public data only.

Why are commit counts sometimes empty? GitHub computes statistics in the background for rarely viewed repos (HTTP 202); they appear on the next run.

Limits

• PR/issue velocity looks at the most recent 100 PRs and issues updated in 30 days; extremely busy repos are sampled.
• Contributor counts follow GitHub's contributors endpoint (top 500 attributed to accounts).

Sources & compliance

• Source: GitHub REST API (api.github.com, version 2022-11-28), read-only GETs, under the GitHub Terms of Service and API terms (https://docs.github.com/en/site-policy/github-terms/github-terms-of-service, section H "API Terms"; checked 11 Oct 2026). Rate limits and headers respected; no scraping of github.com.
• Personal data off by default: contributor usernames are not output unless enabled; no emails are collected. UK GDPR: usernames are public but personal data; enable only if you need them. PECR: no messaging is performed.
• Tokens, webhook URLs and AI keys are secret inputs, redacted from logs and outputs. Webhooks are HTTPS-only and SSRF-guarded.
• Not affiliated with GitHub.