Skip to main content

Free tool

Is this open-source project still healthy? — free health check for any public GitHub repository

Paste a GitHub repository. We read five signals from GitHub's public data, in your browser, and show the verdict with the working.

owner/name, or the address of the repository

Or try one

Nothing you type is sent to us. Your browser asks GitHub directly. A check costs about 7 of the 60 free requests GitHub gives you an hour.

Sample reading · recorded 13 Oct 2026

Healthy

Work lands most weeks, and nothing we read is weak.

frappe/erpnext

CommitsDoes work keep landing?
SteadyStrongLast commit today · a commit in 52 of 52 weeks · 8,094 commits in the year
PeopleDoes one person carry it?
Spread outStrong4 people wrote half of 105 code commits · most: mihir-kandoi (33%)Left out: 2 bot commits, 58 merge commits, 35 docs, chore, CI or test commits
HandlingAre issues and pull requests handled?
Kept upStrongIssues: 931 closed, 498 opened in 90 daysPull requests: merged in a median of 23 hours over 76 merged pull requests from people
RepliesDoes someone answer?
Not read yetPress Look closer to read a dozen recent issuesReads up to 12 recent issues. Costs up to 12 more of your free requests.
ReleasesDoes it ship?
RegularStrongLast release 6 days ago · 30 or more in the year · usually 1 day apart

What to do

Health says nothing about fit. Check the licence, the version you would run, and who would support you.

Recorded from GitHub's public data, so nothing was requested from your browser. Press Check project to read it live, or type or pick another project.

Five signals, each read against a visible line

CHAOSS names these metrics but gives no cut-offs. The lines below are ours: rules of thumb, not tested against projects that later died.

What it reads. Days since the last commit on the default branch, and the weeks with a commit in the last year.

The idea is the OpenSSF Scorecard's: its top score is a commit every week for 90 days. The lines are ours.

Commits · the three lines
Strong Last commit within 30 days, a commit in at least 75% of the last 52 weeks, and no collapse (the last 26 weeks at least 50% of the 26 before)
Fair Last commit within 90 days and a commit in at least 40% of the weeks
Weak Anything slower

frappe/erpnext: Strong

What it reads. The smallest number of people behind half of the code commits, from a sample of up to 200 across the year.

The metric is CHAOSS's: the smallest number of people behind half the contributions. The lines are ours.

People · the three lines
Strong Half the commits come from 4 or more people
Fair Half the commits come from 2 or 3 people
Weak One person wrote more than half

Read only when at least 20 commits count.

In a study of 133 popular projects, 65 % had a truck factor of two or less: the work rests on very few people.

frappe/erpnext: Strong

What it reads. Issues closed per issue opened over 90 days, and the median days to merge a pull request from a person.

CHAOSS defines this ratio for pull requests; we apply it to issues and time pull-request merges instead. Both lines are ours.

Handling · the three lines
Strong Issues closed per issue opened (90 days) at least 0.9, and merged pull requests take 7 days or less (median)
Fair Closed per opened at least 0.5, and a median of 30 days or less to merge
Weak Anything slower; the weaker of the two sides counts

Issues read only when at least 20 were opened; pull requests when 5 merged ones are in the sample.

frappe/erpnext: Strong

What it reads. Of a dozen issues opened 7 to 45 days ago by people outside the team: how many got a team reply, or were closed, within a week.

The metric is CHAOSS's: time to first response, which we sample from a dozen issues instead of measuring all of them. The lines are ours.

Replies · the three lines
Strong At least 70% of a dozen recent issues got a reply from someone on the team, or were closed, within 7 days
Fair At least 40%
Weak Fewer

Read only when at least 8 issues qualify.

frappe/erpnext: not read yet

What it reads. Releases in the last year, days since the last, and the usual gap between them.

The metric is CHAOSS's: the count of releases. The lines are ours.

Releases · the three lines
Strong 4 or more releases in the year, and the last one no older than 1.5 times the usual gap (at least 30 days)
Fair 2 or more in the year, and the last no older than 3 times the usual gap (at least 90 days)
Weak Anything slower, or nothing for 365 days

Not read when the project publishes no GitHub Releases.

frappe/erpnext: Strong

Before any grade, the first of these that applies is the verdict:

  1. Archived or disabled on GitHubArchived
  2. No commits at allNothing to read
  3. Created under 90 days agoToo new
  4. Last commit over 365 days agoDormant
  5. Last commit over 180 days agoStalled
  6. Fewer than 3 signals could be readNot enough to judge

Otherwise it grades: Slowing, Several weak spots, Healthy with things to watch, or Healthy.

What this check answers

Whether an open-source project is still looked after, from its public record.

Your browser asks GitHub directly, and nothing you type is sent to us. A healthy record is not a promise about the future; it is the evidence there is today.

What it rests on

  1. GitHub's public APIReleases, commits, issues and contributors, read in your browser.
  2. The verdictEach signal's reading and threshold are shown under it.

What a good reading does not tell you

Six things the five signals cannot see. Read them before you rely on a verdict.

  1. FitA healthy project can still be wrong for your business.
  2. Quiet is not deadA finished library has few commits and still works. It reads as weak here.
  3. Work done elsewhereForums, ticket systems, GitLab and mailing lists are invisible. So are releases shipped through npm or PyPI.
  4. Money and controlWho pays the maintainers, who owns the company, whether the licence can change: not read.
  5. SecurityHow fast a flaw gets fixed is a different check. The OpenSSF Scorecard reads it.
  6. The futureNothing here is tested against projects that later died.

Who made this

We implement ERPNext and Medusa for clients and support Odoo, and earn from that, so we are not neutral about them. Each is read by the same rules as any other project, and every request is listed so you can check us. We are Frappe specialists; how we stay independent is in this essay.

When the answer is a project

When you decide to run on open source, someone has to look after it. These are the next steps.

  • Talk through your stack

    Bring the projects you checked; we say what we would watch.

  • Managed hosting

    Open-source software on servers you own, updated and watched by us.

  • Our open-source work

    What we have contributed upstream, and where.

Questions about the check

The one most people ask is answered first. Type to find the rest.

Why is there no single score?

A number hides which signal failed. Four of these signals start from CHAOSS's Starter Project Health model, which uses its metrics to improve each project, not to compare projects against each other. You get five readings and one verdict, with the working. To compare two options, check both and read the weak signals, not a total.

The reading

Using it

Start with one conversation

One message, call or WhatsApp is enough. Every line reaches the team directly.

Team away now