diff --git a/util/gather_git_credits.py b/util/gather_git_credits.py
new file mode 100755
index 000000000..1b2865182
--- /dev/null
+++ b/util/gather_git_credits.py
@@ -0,0 +1,67 @@
+#!/usr/bin/env python3
+import subprocess
+import re
+from collections import defaultdict
+
+codefiles = r"(\.[ch](pp)?|\.lua|\.md|\.cmake|\.java|\.gradle|Makefile|CMakeLists\.txt)$"
+
+# two minor versions back, for "Active Contributors"
+REVS_ACTIVE = "5.2.0..HEAD"
+# all time, for "Previous Contributors"
+REVS_PREVIOUS = "HEAD"
+
+CUTOFF_ACTIVE = 3
+CUTOFF_PREVIOUS = 21
+
+# For a description of the points system see:
+# https://github.com/minetest/minetest/pull/9593#issue-398677198
+
+def load(revs):
+	points = defaultdict(int)
+	p = subprocess.Popen(["git", "log", "--mailmap", "--pretty=format:%h %aN <%aE>", revs],
+		stdout=subprocess.PIPE, universal_newlines=True)
+	for line in p.stdout:
+		hash, author = line.strip().split(" ", 1)
+		n = 0
+
+		p2 = subprocess.Popen(["git", "show", "--numstat", "--pretty=format:", hash],
+			stdout=subprocess.PIPE, universal_newlines=True)
+		for line in p2.stdout:
+			added, deleted, filename = re.split(r"\s+", line.strip(), 2)
+			if re.search(codefiles, filename) and added != "-":
+				n += int(added)
+		p2.wait()
+
+		if n == 0:
+			continue
+		if n > 1200:
+			n = 8
+		elif n > 700:
+			n = 4
+		elif n > 100:
+			n = 2
+		else:
+			n = 1
+		points[author] += n
+	p.wait()
+
+	# Some authors duplicate? Don't add manual workarounds here, edit the .mailmap!
+	for author in ("updatepo.sh <script@mt>", "Weblate <42@minetest.ru>"):
+		points.pop(author, None)
+	return points
+
+points_active = load(REVS_ACTIVE)
+points_prev = load(REVS_PREVIOUS)
+
+with open("results.txt", "w") as f:
+	for author, points in sorted(points_active.items(), key=(lambda e: e[1]), reverse=True):
+		if points < CUTOFF_ACTIVE: break
+		points_prev.pop(author, None) # active authors don't appear in previous
+		f.write("%d\t%s\n" % (points, author))
+	f.write('\n---------\n\n')
+	once = True
+	for author, points in sorted(points_prev.items(), key=(lambda e: e[1]), reverse=True):
+		if points < CUTOFF_PREVIOUS and once:
+			f.write('\n---------\n\n')
+			once = False
+		f.write("%d\t%s\n" % (points, author))