diff --git a/src/tools/meson.build b/src/tools/meson.build
index da12d4ac..57aafd5d 100644
--- a/src/tools/meson.build
+++ b/src/tools/meson.build
@@ -14,3 +14,5 @@ executable('wpexec',
   install: true,
   dependencies : [gobject_dep, gio_dep, wp_dep, pipewire_dep],
 )
+
+spa_json_po = files('spa-json-po.py')
diff --git a/src/tools/spa-json-po.py b/src/tools/spa-json-po.py
new file mode 100755
index 00000000..7b1e9dce
--- /dev/null
+++ b/src/tools/spa-json-po.py
@@ -0,0 +1,104 @@
+#!/usr/bin/python3
+# -*- coding: utf-8; mode: python; eval: (blacken-mode); -*-
+"""
+spa-json-po [OPTIONS] FILES...
+
+Extract strings from SPA-JSON files and output a POT file. Values to
+extract are addressed by \"key paths\", such as
+
+    /wireplumber.settings.schema/device.restore-routes/name
+
+which corresponds to the value of the key 'name' in the JSON object.
+Only string values are supported.
+
+"""
+import os
+import sys
+import re
+import argparse
+import subprocess
+import json
+
+
+def main():
+    parser = argparse.ArgumentParser(usage=__doc__.strip())
+    parser.add_argument(
+        "--key-match",
+        "-k",
+        metavar="REGEX",
+        dest="keys",
+        default=[],
+        action="append",
+        help="Regex applied on key path to extract a value",
+    )
+    parser.add_argument(
+        "--output", "-o", default=None, help="Produce output to this file, not stdout"
+    )
+    parser.add_argument(
+        "--spa-json-dump",
+        default="spa-json-dump",
+        help="Path to spa-json-dump executable",
+    )
+    parser.add_argument("files", nargs="+", metavar="FILES", help="Input files")
+    args = parser.parse_args()
+
+    strings = {}
+    for fn in args.files:
+        res = subprocess.run(
+            [args.spa_json_dump, fn], stdout=subprocess.PIPE, check=True
+        )
+        strings.update(parse(res.stdout, keys=args.keys, filename=os.path.basename(fn)))
+
+    if args.output is not None:
+        stream = open(args.output, "w")
+    else:
+        stream = sys.stdout
+
+    try:
+        dump(stream, strings)
+    finally:
+        stream.close()
+
+
+def parse(text, keys, filename):
+    data = json.loads(text)
+    return walk(data, "", keys, filename)
+
+
+def walk(obj, path, keys, filename):
+    result = {}
+
+    if isinstance(obj, str):
+        for key in keys:
+            if re.match(key, path, flags=re.S):
+                result.setdefault(obj, [])
+                result[obj].append((filename, path))
+    elif isinstance(obj, dict):
+        for k, v in obj.items():
+            result.update(walk(v, f"{path}/{k}", keys, filename))
+
+    return result
+
+
+def dump(stream, strings):
+    stream.write(
+        'msgid ""\n'
+        'msgstr ""\n'
+        '"Content-Type: text/plain; charset=UTF-8\\n"\n'
+        '"Content-Transfer-Encoding: 8bit\\n"\n\n'
+    )
+
+    def sort_key(item):
+        msgid, infos = item
+        infos.sort()
+        return infos
+
+    for msgid, infos in sorted(strings.items(), key=sort_key):
+        escaped = json.dumps(msgid)
+        for filename, path in infos:
+            stream.write(f"#. {path}\n#: {filename}\n")
+        stream.write(f'msgid {escaped}\nmsgstr ""\n\n')
+
+
+if __name__ == "__main__":
+    main()