diff --git a/.github/workflows/ci.yaml b/.github/workflows/ci.yaml
new file mode 100644
index 0000000..77055c4
--- /dev/null
+++ b/.github/workflows/ci.yaml
@@ -0,0 +1,19 @@
+name: CI
+
+on: [push]
+
+jobs:
+  ci:
+    name: CI
+    runs-on: ubuntu-latest
+    env:
+      COMPOSE_FILE: docker-compose.yaml
+    steps:
+    - name: Checkout code
+      uses: actions/checkout@v2
+    - name: Build docker image
+      run: make build
+    - name: Lint Check
+      run: make lint
+    - name: Type Check
+      run: make check-type-annotations
diff --git a/Makefile b/Makefile
index 847ce40..fc00683 100644
--- a/Makefile
+++ b/Makefile
@@ -40,15 +40,20 @@ clean:
 
 ## Lint using pylama, isort and black
 lint:
-	$(DOCKER_COMPOSE_RUN) bash -c "pylama && isort --check-only --atomic . && black --check ."
+	$(DOCKER_COMPOSE_RUN) bash -c "pylama && isort --check-only --atomic ./jumpcutter && black --check ./jumpcutter"
 
 ## Sort imports
 sort:
-	$(DOCKER_COMPOSE_RUN) isort --atomic .
+	$(DOCKER_COMPOSE_RUN) isort --atomic ./jumpcutter
 
 ## Format code with black
 format:
-	$(DOCKER_COMPOSE_RUN) black .
+	$(DOCKER_COMPOSE_RUN) black ./jumpcutter
+
+
+## Check type annotations
+check-type-annotations:
+	$(DOCKER_COMPOSE_RUN) mypy ./jumpcutter
 
 ## Lock dependencies with pipenv
 lock_dependencies:
diff --git a/docker-compose.yml b/docker-compose.yaml
similarity index 100%
rename from docker-compose.yml
rename to docker-compose.yaml
diff --git a/jumpcutter/__main__.py b/jumpcutter/__main__.py
index 0885e30..2653f0c 100644
--- a/jumpcutter/__main__.py
+++ b/jumpcutter/__main__.py
@@ -5,7 +5,7 @@
 from jumpcutter.clip import Clip
 
 
-def main():
+def main() -> None:
     parser = argparse.ArgumentParser()
 
     parser.add_argument(
diff --git a/jumpcutter/clip.py b/jumpcutter/clip.py
index 5bfa3b3..bbcc7f3 100644
--- a/jumpcutter/clip.py
+++ b/jumpcutter/clip.py
@@ -1,12 +1,18 @@
+from typing import Dict, List, Tuple
+
 import numpy as np
 
+from moviepy.audio.io.AudioFileClip import AudioFileClip
 from moviepy.editor import VideoFileClip, concatenate_videoclips
 from moviepy.video.fx.all import speedx
+from moviepy.video.io.VideoFileClip import VideoFileClip
 from tqdm import tqdm
 
 
 class Clip:
-    def __init__(self, clip_path, min_loud_part_duration, silence_part_speed):
+    def __init__(
+        self, clip_path: str, min_loud_part_duration: int, silence_part_speed: int
+    ) -> None:
         self.clip = VideoFileClip(clip_path)
         self.audio = Audio(self.clip.audio)
         self.cut_to_method = {
@@ -16,17 +22,14 @@ def __init__(self, clip_path, min_loud_part_duration, silence_part_speed):
         self.min_loud_part_duration = min_loud_part_duration
         self.silence_part_speed = silence_part_speed
 
-    def get_duration(self):
-        return self.clip.duration
-
     def jumpcut(
         self,
-        cuts,
-        magnitude_threshold_ratio,
-        duration_threshold_in_seconds,
-        failure_tolerance_ratio,
-        space_on_edges,
-    ):
+        cuts: List[str],
+        magnitude_threshold_ratio: float,
+        duration_threshold_in_seconds: float,
+        failure_tolerance_ratio: float,
+        space_on_edges: float,
+    ) -> Dict[str, VideoFileClip]:
 
         intervals_to_cut = self.audio.get_intervals_to_cut(
             magnitude_threshold_ratio,
@@ -41,7 +44,9 @@ def jumpcut(
 
         return outputs
 
-    def jumpcut_silent_parts(self, intervals_to_cut):
+    def jumpcut_silent_parts(
+        self, intervals_to_cut: List[Tuple[float, float]]
+    ) -> List[VideoFileClip]:
         jumpcutted_clips = []
         previous_stop = 0
         for start, stop in tqdm(intervals_to_cut, desc="Cutting silent intervals"):
@@ -64,7 +69,9 @@ def jumpcut_silent_parts(self, intervals_to_cut):
             jumpcutted_clips.append(last_clip)
         return jumpcutted_clips
 
-    def jumpcut_voiced_parts(self, intervals_to_cut):
+    def jumpcut_voiced_parts(
+        self, intervals_to_cut: List[Tuple[float, float]]
+    ) -> List[VideoFileClip]:
         jumpcutted_clips = []
         for start, stop in tqdm(intervals_to_cut, desc="Cutting voiced intervals"):
             if start < stop:
@@ -74,7 +81,7 @@ def jumpcut_voiced_parts(self, intervals_to_cut):
 
 
 class Audio:
-    def __init__(self, audio):
+    def __init__(self, audio: AudioFileClip) -> None:
         self.audio = audio
         self.fps = audio.fps
 
@@ -84,11 +91,11 @@ def __init__(self, audio):
 
     def get_intervals_to_cut(
         self,
-        magnitude_threshold_ratio,
-        duration_threshold_in_seconds,
-        failure_tolerance_ratio,
-        space_on_edges,
-    ):
+        magnitude_threshold_ratio: float,
+        duration_threshold_in_seconds: float,
+        failure_tolerance_ratio: float,
+        space_on_edges: float,
+    ) -> List[Tuple[float, float]]:
         min_magnitude = min(abs(np.min(self.signal)), np.max(self.signal))
         magnitude_threshold = min_magnitude * magnitude_threshold_ratio
         failure_tolerance = self.fps * failure_tolerance_ratio
diff --git a/setup.cfg b/setup.cfg
index 0b6d346..df7bf30 100644
--- a/setup.cfg
+++ b/setup.cfg
@@ -24,7 +24,7 @@ show_closures = True
 no_assert = True
 
 [mypy]
-python_version = 3.6
+python_version = 3.8
 warn_redundant_casts = False
 warn_unused_ignores = True
 warn_unused_configs = True