diff --git a/backend/donations/extract.py b/backend/donations/extract.py
new file mode 100644
index 00000000..f44d0617
--- /dev/null
+++ b/backend/donations/extract.py
@@ -0,0 +1,66 @@
+"""
+Extract the CNP and Address from the PDF form
+
+TODO: Integrate this with Django as a management command and use Django's FileField
+"""
+
+from collections import namedtuple
+from functools import partial
+from io import BytesIO
+from pypdf import PdfReader
+
+
+Zone = namedtuple("Zone", ["start_y", "start_x", "end_x"])
+
+
+data_zones = {
+    "cnp": Zone(671, 336, 0),
+    "father": Zone(681, 300, 336),
+    "street_name": Zone(636, 67, 289),
+    "street_number": Zone(636, 289, 368),
+    "street_bl": Zone(614, 49, 108),
+    "street_sc": Zone(614, 108, 150),
+    "street_et": Zone(614, 150, 185),
+    "street_ap": Zone(614, 185, 255),
+    "percent": Zone(341, 146, 186),
+}
+
+
+def _visitor_builder(parts, start_x, start_y, end_x, text, cm, tm, fontDict, fontSize):
+    # print(tm, text)  # Print the text matrix (for debugging)
+    x = tm[4]
+    y = tm[5]
+    # Check that the text is on the required line
+    if x >= start_x and y == start_y:
+        # Check that the text doesn't run over the end position
+        if (end_x and x < end_x) or not end_x:
+            parts.append(text.strip())
+
+
+def extract_data(page, zone: Zone):
+    parts = []
+    visitor = partial(_visitor_builder, parts, zone.start_x, zone.start_y, zone.end_x)
+    page.extract_text(0, extraction_mode="plain", visitor_text=visitor)
+    return "".join(parts)
+
+
+if __name__ == "__main__":
+    # This "open" to be replaced with Django's FileField.open()
+    pdf_file = open("document.pdf", "rb")
+
+    reader = PdfReader(BytesIO(pdf_file.read()))
+    page = reader.pages[0]
+
+    print("CNP =", extract_data(page, data_zones["cnp"]))
+    print("Father =", extract_data(page, data_zones["father"]))
+
+    print("Street Name =", extract_data(page, data_zones["street_name"]))
+    print("Street Number =", extract_data(page, data_zones["street_number"]))
+    print("BL =", extract_data(page, data_zones["street_bl"]))
+    print("SC =", extract_data(page, data_zones["street_sc"]))
+    print("ET =", extract_data(page, data_zones["street_et"]))
+    print("AP =", extract_data(page, data_zones["street_ap"]))
+
+    print("Percent =", extract_data(page, data_zones["percent"]))
+
+    pdf_file.close()