diff --git a/personal_data/fetchers/home_assistant.py b/personal_data/fetchers/home_assistant.py
new file mode 100644
index 0000000..109c0bc
--- /dev/null
+++ b/personal_data/fetchers/home_assistant.py
@@ -0,0 +1,52 @@
+import dataclasses
+from decimal import Decimal
+import datetime
+import logging
+from collections.abc import Iterator, Mapping
+
+from frozendict import frozendict
+
+from personal_data.data import DeduplicateMode, Scraper
+
+from .. import secrets
+
+logger = logging.getLogger(__name__)
+
+HA_ROOT = secrets.HOME_ASSISTANT_ROOT
+HA_LLAK= secrets.HOME_ASSISTANT_LLAK
+
+@dataclasses.dataclass(frozen=True)
+class HomeAssistantScaleWeight(Scraper):
+    dataset_name = 'health_weight'
+    deduplicate_mode = DeduplicateMode.BY_ALL_COLUMNS
+    deduplicate_ignore_columns = []
+
+    def scrape(self) -> Iterator[Mapping[str, object]]:
+        headers = {
+                'Authorization': 'Bearer '+HA_LLAK,
+                'Content-Type': 'application/json',
+        }
+        end_time = datetime.datetime.now()
+        start_time = end_time - datetime.timedelta(days = 90)
+        url = '{}/api/history/period/{}'.format(HA_ROOT, start_time)
+        print(url)
+        params = {
+                'filter_entity_id': 'sensor.bathroom_scale_mass',
+                'end_time': end_time,
+        }
+        response = self.session.get(url, params = params, headers = headers)
+        response.raise_for_status()
+
+        data = response.json()
+
+        state_range_for_consideration = (Decimal(30), Decimal(300))
+
+        for d in data[0]:
+            if d['state'] == 'unavailable':
+                continue
+            state = Decimal(d['state'])
+            if state_range_for_consideration[0] <= state <= state_range_for_consideration[1]:
+                yield {
+                        'weight.sample_time': datetime.datetime.fromisoformat(d['last_updated']),
+                        'weight.kg': state,
+                }
diff --git a/personal_data/main.py b/personal_data/main.py
index 949fcdc..0e2b2c3 100644
--- a/personal_data/main.py
+++ b/personal_data/main.py
@@ -77,6 +77,24 @@ def equals_without_fields(
 
     return frozendict(a) == frozendict(b)
 
+def deduplicate_by_ignoring_certain_fields(dicts: list[dict],
+    deduplicate_ignore_columns: list[str],
+) -> list[dict]:
+    """Removes duplicates that occur when ignoring certain columns.
+
+    Output order is stable.
+    """
+    to_remove = set()
+    for idx1, first in enumerate(dicts):
+        for idx2, second in enumerate(dicts[idx1 + 1 :], idx1 + 1):
+            if equals_without_fields(first, second, deduplicate_ignore_columns):
+                to_remove.add(idx2)
+
+    to_remove = sorted(to_remove)
+    while to_remove:
+        del dicts[to_remove.pop()]
+
+    return dicts
 
 def deduplicate_dicts(
     dicts: Sequence[dict],
@@ -99,12 +117,8 @@ def deduplicate_dicts(
         ):
             del dicts[-1]
     elif deduplicate_mode == personal_data.data.DeduplicateMode.BY_ALL_COLUMNS:
-        to_remove = set()
-        for idx1, first in enumerate(dicts):
-            for second in dicts[idx1 + 1 :]:
-                if equals_without_fields(first, second, deduplicate_ignore_columns):
-                    to_remove.add(second)
-        dicts = set(dicts) - to_remove
+        dicts = deduplicate_by_ignoring_certain_fields(dicts,
+                                                       deduplicate_ignore_columns)
     elif deduplicate_mode != personal_data.data.DeduplicateMode.NONE:
         dicts = set(dicts)
 
diff --git a/personal_data/secrets.py b/personal_data/secrets.py
index a68bd76..7984c1e 100644
--- a/personal_data/secrets.py
+++ b/personal_data/secrets.py
@@ -43,6 +43,10 @@ KUCOIN_PASS = load_secret('KUCOIN_PASS')
 KRAKEN_KEY = load_secret('KRAKEN_KEY')
 KRAKEN_SECRET = load_secret('KRAKEN_SECRET')
 
+# Home Assistant
+HOME_ASSISTANT_ROOT = load_secret('HOME_ASSISTANT_ROOT')
+HOME_ASSISTANT_LLAK = load_secret('HOME_ASSISTANT_LLAK')
+
 # Email configuration
 MAILGUN_API_KEY = load_secret('MAILGUN_API_KEY')
 MAILGUN_DOMAIN = load_secret('MAILGUN_DOMAIN')
diff --git a/test/test_deduplicate.py b/test/test_deduplicate.py
index 55729d7..f103097 100644
--- a/test/test_deduplicate.py
+++ b/test/test_deduplicate.py
@@ -1,5 +1,8 @@
 from frozendict import frozendict
 
+import datetime
+from decimal import Decimal
+
 from personal_data.data import DeduplicateMode
 from personal_data.main import deduplicate_dicts
 
@@ -23,7 +26,7 @@ def test_no_fields_to_ignore():
 def test_only_latest():
     ls, fields = deduplicate_dicts(LIST, DeduplicateMode.ONLY_LATEST, ['t'])
     assert fields == ['a', 'b', 't']
-    assert ls == ls[:3]
+    assert ls == LIST[:3]
 
 
 def test_all_fields():
@@ -34,3 +37,28 @@ def test_all_fields():
         frozendict({'a': 0, 'b': 12, 't': 300}),
         frozendict({'a': 1, 'b': 2, 't': 300}),
     ]
+
+def test_all_fields():
+    ls, fields = deduplicate_dicts(LIST + LIST, DeduplicateMode.BY_ALL_COLUMNS, ['t'])
+    assert fields == ['a', 'b', 't']
+    print(ls)
+    assert ls == [
+        frozendict({'a': 0, 'b': 12, 't': 300}),
+        frozendict({'a': 1, 'b': 2, 't': 300}),
+    ]
+
+LIST_2 = [
+        frozendict({'weight.sample_time': datetime.datetime(2024, 5, 28, 6, 27, 31, 134506, tzinfo=datetime.timezone.utc), 'weight.kg': Decimal('73.6')}),
+        frozendict({'weight.sample_time': datetime.datetime(2024, 6, 1, 7, 36, 9, 590355, tzinfo=datetime.timezone.utc), 'weight.kg': Decimal('74.7')}),
+]
+
+def test_deduplicate_weight():
+    ls, fields = deduplicate_dicts(LIST_2, DeduplicateMode.BY_ALL_COLUMNS, [])
+    assert fields == ['weight.sample_time', 'weight.kg']
+    assert ls == LIST_2
+
+def test_deduplicate_weight_2():
+    ls, fields = deduplicate_dicts(LIST_2 + LIST_2, DeduplicateMode.BY_ALL_COLUMNS, [])
+    assert fields == ['weight.sample_time', 'weight.kg']
+    assert ls == LIST_2
+