diff --git a/.gitignore b/.gitignore
--- a/.gitignore
+++ b/.gitignore
@@ -9,3 +9,4 @@
 dist/
 version.txt
 .tox
+.mypy_cache/
diff --git a/MANIFEST.in b/MANIFEST.in
--- a/MANIFEST.in
+++ b/MANIFEST.in
@@ -3,3 +3,4 @@
 include requirements-swh.txt
 include version.txt
 include README.md
+recursive-include swh py.typed
diff --git a/mypy.ini b/mypy.ini
new file mode 100644
--- /dev/null
+++ b/mypy.ini
@@ -0,0 +1,18 @@
+[mypy]
+namespace_packages = True
+warn_unused_ignores = True
+
+
+# 3rd party libraries without stubs (yet)
+
+[mypy-elasticsearch.*]
+ignore_missing_imports = True
+
+[mypy-msgpack.*]
+ignore_missing_imports = True
+
+[mypy-pkg_resources.*]
+ignore_missing_imports = True
+
+[mypy-pytest.*]
+ignore_missing_imports = True
diff --git a/swh/__init__.py b/swh/__init__.py
--- a/swh/__init__.py
+++ b/swh/__init__.py
@@ -1 +1,4 @@
-__path__ = __import__('pkgutil').extend_path(__path__, __name__)
+from pkgutil import extend_path
+from typing import Iterable
+
+__path__ = extend_path(__path__, __name__)  # type: Iterable[str]
diff --git a/swh/search/cli.py b/swh/search/cli.py
--- a/swh/search/cli.py
+++ b/swh/search/cli.py
@@ -73,7 +73,7 @@
 
 
 @cli.command('rpc-serve')
-@click.argument('config-path', required=1)
+@click.argument('config-path', required=True)
 @click.option('--host', default='0.0.0.0', help="Host to run the server")
 @click.option('--port', default=5010, type=click.INT,
               help="Binding port of the server")
diff --git a/swh/search/elasticsearch.py b/swh/search/elasticsearch.py
--- a/swh/search/elasticsearch.py
+++ b/swh/search/elasticsearch.py
@@ -4,7 +4,7 @@
 # See top-level LICENSE file for more information
 
 import base64
-from typing import Iterable, Dict, List, Iterator
+from typing import Any, Iterable, Dict, List, Iterator, Optional
 
 from elasticsearch import Elasticsearch
 from elasticsearch.helpers import bulk, scan
@@ -131,7 +131,7 @@
               list of dictionaries with key:
               * `url`: URL of a matching origin
         """
-        query_clauses = []
+        query_clauses = []  # type: List[Dict[str, Any]]
 
         if url_pattern:
             query_clauses.append({
@@ -185,9 +185,11 @@
         }
         if scroll_token:
             # TODO: use ElasticSearch's scroll API?
-            scroll_token = msgpack.loads(base64.b64decode(scroll_token))
+            scroll_token_content = msgpack.loads(
+                base64.b64decode(scroll_token))
             body['search_after'] = \
-                [scroll_token[b'score'], scroll_token[b'id'].decode('ascii')]
+                [scroll_token_content[b'score'],
+                 scroll_token_content[b'id'].decode('ascii')]
 
         res = self._backend.search(
             index='origin',
@@ -199,12 +201,12 @@
 
         if len(hits) == count:
             last_hit = hits[-1]
-            next_scroll_token = {
+            next_scroll_token_content = {
                 b'score': last_hit['_score'],
                 b'id': last_hit['_id'],
             }
             next_scroll_token = base64.b64encode(msgpack.dumps(
-                next_scroll_token))
+                next_scroll_token_content))  # type: Optional[bytes]
         else:
             next_scroll_token = None
 
diff --git a/swh/search/in_memory.py b/swh/search/in_memory.py
--- a/swh/search/in_memory.py
+++ b/swh/search/in_memory.py
@@ -7,7 +7,7 @@
 from collections import defaultdict
 import itertools
 import re
-from typing import Iterable, Dict
+from typing import Any, Dict, Iterable, Iterator, List, Optional
 
 import msgpack
 
@@ -40,8 +40,8 @@
             del self._origin_ids
 
     def initialize(self) -> None:
-        self._origins = defaultdict(dict)
-        self._origin_ids = []
+        self._origins = defaultdict(dict)  # type: Dict[str, Dict[str, Any]]
+        self._origin_ids = []  # type: List[str]
 
     _url_splitter = re.compile(r'\W')
 
@@ -64,7 +64,7 @@
             with_visit: bool = False,
             scroll_token: str = None, count: int = 50
             ) -> Dict[str, object]:
-        matches = (self._origins[id_] for id_ in self._origin_ids)
+        matches = (self._origins[id_] for id_ in self._origin_ids)  # type: Iterator[Dict[str, Any]]
 
         if url_pattern:
             tokens = set(self._url_splitter.split(url_pattern))
@@ -96,8 +96,9 @@
             matches = filter(lambda o: o.get('has_visits'), matches)
 
         if scroll_token:
-            scroll_token = msgpack.loads(base64.b64decode(scroll_token))
-            start_at_index = scroll_token[b'start_at_index']
+            scroll_token_content = msgpack.loads(
+                base64.b64decode(scroll_token))
+            start_at_index = scroll_token_content[b'start_at_index']
         else:
             start_at_index = 0
 
@@ -105,11 +106,11 @@
             matches, start_at_index, start_at_index+count))
 
         if len(hits) == count:
-            next_scroll_token = {
+            next_scroll_token_content = {
                 b'start_at_index': start_at_index+count,
             }
             next_scroll_token = base64.b64encode(msgpack.dumps(
-                next_scroll_token))
+                next_scroll_token_content))  # type: Optional[bytes]
         else:
             next_scroll_token = None
 
diff --git a/swh/search/py.typed b/swh/search/py.typed
new file mode 100644
--- /dev/null
+++ b/swh/search/py.typed
@@ -0,0 +1 @@
+# Marker file for PEP 561.
diff --git a/tox.ini b/tox.ini
--- a/tox.ini
+++ b/tox.ini
@@ -1,5 +1,5 @@
 [tox]
-envlist=flake8,py3
+envlist=flake8,mypy,py3
 
 [testenv:py3]
 deps =
@@ -16,3 +16,11 @@
   flake8
 commands =
   {envpython} -m flake8
+
+[testenv:mypy]
+skip_install = true
+deps =
+  .[testing]
+  mypy
+commands =
+  mypy swh