diff --git a/wsgi.py b/wsgi.py
index cc00218..2e03977 100644
--- a/wsgi.py
+++ b/wsgi.py
@@ -2,8 +2,37 @@
 Point d'entrée Gunicorn pour la production.
 Usage : gunicorn -c gunicorn.conf.py wsgi:app
 """
-from app import app, _init_db, start_scheduler
+from app import app, db, _init_db, start_scheduler
+from app import CrawlSource
+
+_DEFAULT_SOURCES = [
+    ('HAS',                       'https://www.has-sante.fr/',                                                                           'webpage', 'scientific'),
+    ('Ameli',                     'https://www.ameli.fr/',                                                                               'index',   'advice'),
+    ('Sante publique France',     'https://www.santepubliquefrance.fr/',                                                                 'index',   'advice'),
+    ('Top sante',                 'https://www.topsante.com/',                                                                           'index',   'advice'),
+    ('Examine.com - Supplements', 'https://examine.com/supplements/',                                                                    'index',   'scientific'),
+    ('Examine.com - Nutrition',   'https://examine.com/nutrition/',                                                                      'index',   'scientific'),
+    ('NutritionFacts.org',        'https://nutritionfacts.org/feed/',                                                                    'rss',     'advice'),
+    ('Harvard Health Blog',       'https://www.health.harvard.edu/blog/feed',                                                            'rss',     'advice'),
+    ('ScienceDaily - Sante',      'https://www.sciencedaily.com/rss/health_medicine.xml',                                               'rss',     'advice'),
+    ('Medical News Today',        'https://www.medicalnewstoday.com/rss',                                                                'rss',     'advice'),
+    ('NEJM - Current Issue',      'https://www.nejm.org/action/showFeed?jc=nejm&type=etoc&feed=rss',                                    'rss',     'scientific'),
+    ('BMJ - Current Issue',       'https://www.bmj.com/rss/thebmj.xml',                                                                 'rss',     'scientific'),
+    ('The Lancet',                'https://www.thelancet.com/rssfeed/lancet_current.xml',                                               'rss',     'scientific'),
+    ('PubMed - Nutrition',        'https://pubmed.ncbi.nlm.nih.gov/rss/search/0f8VGqGLrpuLRhtqFbgA5gf7wSDBrw_V/?limit=20&format=abstract', 'rss', 'scientific'),
+]
+
+
+def _seed_sources():
+    """Insert default sources if the table is empty."""
+    if CrawlSource.query.count() > 0:
+        return
+    for name, url, stype, cat in _DEFAULT_SOURCES:
+        db.session.add(CrawlSource(name=name, url=url, source_type=stype, category=cat))
+    db.session.commit()
+
 
 with app.app_context():
     _init_db()
+    _seed_sources()
     start_scheduler()