diff --git a/app/scodoc/sco_trombino.py b/app/scodoc/sco_trombino.py
index e9102fc67..9534cfb29 100644
--- a/app/scodoc/sco_trombino.py
+++ b/app/scodoc/sco_trombino.py
@@ -55,19 +55,18 @@ from app.scodoc.sco_pdf import SU
 from app.scodoc import html_sco_header
 from app.scodoc import htmlutils
 from app.scodoc import sco_import_etuds
+from app.scodoc import sco_etud
 from app.scodoc import sco_excel
-from app.scodoc import sco_formsemestre
-from app.scodoc import sco_groups
 from app.scodoc import sco_groups_view
 from app.scodoc import sco_pdf
 from app.scodoc import sco_photos
 from app.scodoc import sco_portal_apogee
 from app.scodoc import sco_preferences
-from app.scodoc import sco_etud
+from app.scodoc import sco_trombino_doc
 
 
 def trombino(
-    group_ids=[],  # liste des groupes à afficher
+    group_ids=(),  # liste des groupes à afficher
     formsemestre_id=None,  # utilisé si pas de groupes selectionné
     etat=None,
     format="html",
@@ -93,6 +92,8 @@ def trombino(
         return _trombino_pdf(groups_infos)
     elif format == "pdflist":
         return _listeappel_photos_pdf(groups_infos)
+    elif format == "doc":
+        return sco_trombino_doc.trombino_doc(groups_infos)
     else:
         raise Exception("invalid format")
         # return _trombino_html_header() + trombino_html( group, members) + html_sco_header.sco_footer()
@@ -176,8 +177,13 @@ def trombino_html(groups_infos):
 
     H.append("</div>")
     H.append(
-        '<div style="margin-bottom:15px;"><a class="stdlink" href="trombino?format=pdf&%s">Version PDF</a></div>'
-        % groups_infos.groups_query_args
+        f"""<div style="margin-bottom:15px;">
+        <a class="stdlink" href="{url_for('scolar.trombino', scodoc_dept=g.scodoc_dept, 
+            format='pdf', group_ids=groups_infos.group_ids)}">Version PDF</a>
+        &nbsp;&nbsp;
+        <a class="stdlink" href="{url_for('scolar.trombino', scodoc_dept=g.scodoc_dept, 
+            format='doc', group_ids=groups_infos.group_ids)}">Version doc</a>
+        </div>"""
     )
     return "\n".join(H)
 
@@ -234,7 +240,7 @@ def _trombino_zip(groups_infos):
         Z.writestr(filename, img)
     Z.close()
     size = data.tell()
-    log("trombino_zip: %d bytes" % size)
+    log(f"trombino_zip: {size} bytes")
     data.seek(0)
     return send_file(
         data,
@@ -470,7 +476,7 @@ def _listeappel_photos_pdf(groups_infos):
 
 
 # ---------------------    Upload des photos de tout un groupe
-def photos_generate_excel_sample(group_ids=[]):
+def photos_generate_excel_sample(group_ids=()):
     """Feuille excel pour import fichiers photos"""
     fmt = sco_import_etuds.sco_import_format()
     data = sco_import_etuds.sco_import_generate_excel_sample(
@@ -492,31 +498,33 @@ def photos_generate_excel_sample(group_ids=[]):
     # return sco_excel.send_excel_file(data, "ImportPhotos" + scu.XLSX_SUFFIX)
 
 
-def photos_import_files_form(group_ids=[]):
+def photos_import_files_form(group_ids=()):
     """Formulaire pour importation photos"""
     if not group_ids:
         raise ScoValueError("paramètre manquant !")
     groups_infos = sco_groups_view.DisplayedGroupsInfos(group_ids)
-    back_url = "groups_view?%s&curtab=tab-photos" % groups_infos.groups_query_args
+    back_url = f"groups_view?{groups_infos.groups_query_args}&curtab=tab-photos"
 
     H = [
         html_sco_header.sco_header(page_title="Import des photos des étudiants"),
-        """<h2 class="formsemestre">Téléchargement des photos des étudiants</h2>
-         <p><b>Vous pouvez aussi charger les photos individuellement via la fiche de chaque étudiant (menu "Etudiant" / "Changer la photo").</b></p>
-         <p class="help">Cette page permet de charger en une seule fois les photos de plusieurs étudiants.<br/>
-          Il faut d'abord remplir une feuille excel donnant les noms 
+        f"""<h2 class="formsemestre">Téléchargement des photos des étudiants</h2>
+         <p><b>Vous pouvez aussi charger les photos individuellement via la fiche
+         de chaque étudiant (menu "Etudiant" / "Changer la photo").</b>
+         </p>
+         <p class="help">Cette page permet de charger en une seule fois les photos
+         de plusieurs étudiants.<br/>
+          Il faut d'abord remplir une feuille excel donnant les noms
           des fichiers images (une image par étudiant).
          </p>
-         <p class="help">Ensuite, réunir vos images dans un fichier zip, puis télécharger 
+         <p class="help">Ensuite, réunir vos images dans un fichier zip, puis télécharger
          simultanément le fichier excel et le fichier zip.
          </p>
         <ol>
-        <li><a class="stdlink" href="photos_generate_excel_sample?%s">
+        <li><a class="stdlink" href="photos_generate_excel_sample?{groups_infos.groups_query_args}">
         Obtenir la feuille excel à remplir</a>
         </li>
         <li style="padding-top: 2em;">
-         """
-        % groups_infos.groups_query_args,
+         """,
     ]
     F = html_sco_header.sco_footer()
     vals = scu.get_request_args()
diff --git a/app/scodoc/sco_trombino_doc.py b/app/scodoc/sco_trombino_doc.py
new file mode 100644
index 000000000..038832e5a
--- /dev/null
+++ b/app/scodoc/sco_trombino_doc.py
@@ -0,0 +1,76 @@
+##############################################################################
+# ScoDoc
+# Copyright (c) 1999 - 2022 Emmanuel Viennet.  All rights reserved.
+# See LICENSE
+##############################################################################
+
+"""Génération d'un trombinoscope en doc
+"""
+
+import docx
+from docx.shared import Mm
+from docx.enum.text import WD_ALIGN_PARAGRAPH
+from docx.enum.table import WD_ALIGN_VERTICAL
+
+from app.scodoc import sco_etud
+from app.scodoc import sco_photos
+import app.scodoc.sco_utils as scu
+import sco_version
+
+
+def trombino_doc(groups_infos):
+    "Send photos as docx document"
+    filename = f"trombino_{groups_infos.groups_filename}.docx"
+    sem = groups_infos.formsemestre  # suppose 1 seul semestre
+    PHOTO_WIDTH = Mm(25)
+    N_PER_ROW = 5  # XXX should be in ScoDoc preferences
+
+    document = docx.Document()
+    document.add_heading(
+        f"Trombinoscope {sem['titreannee']} {groups_infos.groups_titles}", 1
+    )
+    section = document.sections[0]
+    footer = section.footer
+    footer.paragraphs[
+        0
+    ].text = f"Généré par {sco_version.SCONAME} le {scu.timedate_human_repr()}"
+
+    nb_images = len(groups_infos.members)
+    table = document.add_table(rows=2 * (nb_images // N_PER_ROW + 1), cols=N_PER_ROW)
+    table.allow_autofit = False
+
+    for i, t in enumerate(groups_infos.members):
+        li = i // N_PER_ROW
+        co = i % N_PER_ROW
+        img_path = (
+            sco_photos.photo_pathname(t["photo_filename"], size="small")
+            or sco_photos.UNKNOWN_IMAGE_PATH
+        )
+        cell = table.rows[2 * li].cells[co]
+        cell.vertical_alignment = WD_ALIGN_VERTICAL.TOP
+        cell_p, cell_f, cell_r = _paragraph_format_run(cell)
+        cell_r.add_picture(img_path, width=PHOTO_WIDTH)
+
+        # le nom de l'étudiant: cellules de lignes impaires
+        cell = table.rows[2 * li + 1].cells[co]
+        cell.vertical_alignment = WD_ALIGN_VERTICAL.TOP
+        cell_p, cell_f, cell_r = _paragraph_format_run(cell)
+        cell_r.add_text(sco_etud.format_nomprenom(t))
+        cell_f.space_after = Mm(8)
+
+    return scu.send_docx(document, filename)
+
+
+def _paragraph_format_run(cell):
+    "parag. dans cellule tableau"
+    # inspired by https://stackoverflow.com/questions/64218305/problem-with-python-docx-putting-pictures-in-a-table
+    paragraph = cell.paragraphs[0]
+    fmt = paragraph.paragraph_format
+    run = paragraph.add_run()
+
+    fmt.space_before = Mm(0)
+    fmt.space_after = Mm(0)
+    fmt.line_spacing = 1.0
+    fmt.alignment = WD_ALIGN_PARAGRAPH.CENTER
+
+    return paragraph, fmt, run
diff --git a/app/scodoc/sco_utils.py b/app/scodoc/sco_utils.py
index 0a4da781a..8cba3a59c 100644
--- a/app/scodoc/sco_utils.py
+++ b/app/scodoc/sco_utils.py
@@ -33,6 +33,7 @@ import bisect
 import copy
 import datetime
 from enum import IntEnum
+import io
 import json
 from hashlib import md5
 import numbers
@@ -49,6 +50,7 @@ from PIL import Image as PILImage
 import pydot
 import requests
 
+import flask
 from flask import g, request
 from flask import flash, url_for, make_response, jsonify
 
@@ -379,6 +381,10 @@ CSV_FIELDSEP = ";"
 CSV_LINESEP = "\n"
 CSV_MIMETYPE = "text/comma-separated-values"
 CSV_SUFFIX = ".csv"
+DOCX_MIMETYPE = (
+    "application/vnd.openxmlformats-officedocument.wordprocessingml.document"
+)
+DOCX_SUFFIX = ".docx"
 JSON_MIMETYPE = "application/json"
 JSON_SUFFIX = ".json"
 PDF_MIMETYPE = "application/pdf"
@@ -398,6 +404,7 @@ def get_mime_suffix(format_code: str) -> tuple[str, str]:
     """
     d = {
         "csv": (CSV_MIMETYPE, CSV_SUFFIX),
+        "docx": (DOCX_MIMETYPE, DOCX_SUFFIX),
         "xls": (XLSX_MIMETYPE, XLSX_SUFFIX),
         "xlsx": (XLSX_MIMETYPE, XLSX_SUFFIX),
         "pdf": (PDF_MIMETYPE, PDF_SUFFIX),
@@ -740,6 +747,18 @@ def send_file(data, filename="", suffix="", mime=None, attached=None):
     return response
 
 
+def send_docx(document, filename):
+    "Send a python-docx document"
+    buffer = io.BytesIO()  # in-memory document, no disk file
+    document.save(buffer)
+    buffer.seek(0)
+    return flask.send_file(
+        buffer,
+        attachment_filename=sanitize_filename(filename),
+        mimetype=DOCX_MIMETYPE,
+    )
+
+
 def get_request_args():
     """returns a dict with request (POST or GET) arguments
     converted to suit legacy Zope style (scodoc7) functions.