diff --git a/.github/workflows/tests.yml b/.github/workflows/tests.yml
new file mode 100644
index 0000000..a1cb1d1
--- /dev/null
+++ b/.github/workflows/tests.yml
@@ -0,0 +1,46 @@
+# test based on ChatGPT's suggestion
+name: CI
+
+on:
+  push:
+    branches: [ master ]                # run for every push to master
+  pull_request:                         # run for every PR
+
+jobs:
+  tests:
+    runs-on: ubuntu-latest              # GitHub-hosted Linux runner
+
+    steps:
+      # 1️⃣  Bring the source onto the runner
+      - name: Check out code
+        uses: actions/checkout@v4
+
+      # 2️⃣  System packages for build & tests
+      - name: Install build dependencies
+        run: |
+          sudo apt-get update
+          sudo apt-get install -y \
+              build-essential gfortran make \
+              libfftw3-dev liblapack-dev \
+              python3 python3-numpy
+
+      # 3️⃣  Compile the QDyn executable (bin ends up in src/qdyn)
+      - name: Build qdyn
+        run: |
+          make -C src clean
+          make -C src \
+            LIBS="-lfftw3 -lm -llapack"   # override local path in Makefile
+
+      # 4️⃣  Run the full regression-test suite
+      - name: Run test suite
+        run: |
+          cd tests
+          bash run_test_suite.sh | tee ../test_output.log
+
+      # 5️⃣  Fail the workflow if any test reported an error
+      - name: Check for test failures
+        run: |
+          if grep "ERROR found" test_output.log; then
+            echo "Some tests failed — see log above."
+            exit 1
+          fi
diff --git a/src/Makefile b/src/Makefile
index 3720cf7..1071528 100644
--- a/src/Makefile
+++ b/src/Makefile
@@ -3,9 +3,8 @@
 # Compiler
 FC = gfortran
 FFLAGS =
-# FFTW library needed
-LIBS = -L/Users/janosj/Documents/Programs/fftw/fftw-3.3.10/lib -lfftw3 -lm -llapack
-LDLIBS = -lfftw3 # This is necessary only on clusters, not on MacOS
+LIBS = -L/Users/janosj/Documents/Programs/fftw/fftw-3.3.10/lib -lfftw3 -lm -llapack # FFTW and lapack libraries needed
+LDLIBS =
 
 # Name of program
 OUT = qdyn
@@ -16,7 +15,7 @@ OUT = qdyn
 OBJS = vars.o fparser.o fftw.o utils.o exactfactor.o init.o propag.o qdyn.o
 
 myprogram: $(OBJS) 
-	   $(FC) -o $(OUT) ${LIBS} ${FFLAGS} $(OBJS) ${LDLIBS}
+	   $(FC) -o $(OUT) ${FFLAGS} $(OBJS) ${LIBS} ${LDLIBS}
 
 clean:
 	/bin/rm -f *.o *.mod $(OUT)
diff --git a/tests/EF_LaserPulse_OH_1D/check_ef.py b/tests/EF_LaserPulse_OH_1D/check_ef.py
index ce0ed8b..c6bbb8b 100644
--- a/tests/EF_LaserPulse_OH_1D/check_ef.py
+++ b/tests/EF_LaserPulse_OH_1D/check_ef.py
@@ -26,8 +26,8 @@
     # calculate maximum difference in the files (including x axis which shouldn't change)
     max_difference = max(abs(data - ref))
 
-    # compare max difference to threshold
-    if max_difference > 1e-10:
+    # compare max difference to threshold (set based on difference produced by my Mac and GitHub Actions)
+    if max_difference > 5e-8:
         result = False
         output.write(f"Mismatch found in '{data_file}'! (max difference: {max_difference:.2e})\n")
 
diff --git a/tests/RT_LaserPulse_CH3I_1D/check_ch3i.py b/tests/RT_LaserPulse_CH3I_1D/check_ch3i.py
index b349f06..9ead1ed 100644
--- a/tests/RT_LaserPulse_CH3I_1D/check_ch3i.py
+++ b/tests/RT_LaserPulse_CH3I_1D/check_ch3i.py
@@ -1,13 +1,12 @@
 from os.path import exists
 
-import matplotlib.pyplot as plt
 from numpy import genfromtxt, linspace, interp, std, trapz, sqrt, array, reshape, shape
 
 xngrid = 256
 
 # reading calculated data
 wf = genfromtxt('wf1d_ad.1.dat')
-nframes_wf = int(shape(wf)[0]/xngrid)
+nframes_wf = int(shape(wf)[0] / xngrid)
 wf = reshape(wf, (nframes_wf, xngrid, 5)).transpose((0, 2, 1))
 energy = genfromtxt('energies.dat').T
 t = energy[0]
@@ -33,12 +32,12 @@
 # calculating mean <x> and <delta x>
 x = [[], [], []]
 for i in range(0, nframes_wf, 10):
-    x[0].append(t[i]*0.02418884254)
+    x[0].append(t[i] * 0.02418884254)
     dx = wf[i][0][1] - wf[i][0][1]
     norm = trapz(wf[i][3], wf[i][0], dx=dx)
-    aver_x = trapz(wf[i][3]*wf[i][0], wf[i][0], dx=dx)/norm
+    aver_x = trapz(wf[i][3] * wf[i][0], wf[i][0], dx=dx) / norm
     x[1].append(aver_x)
-    aver_dx = sqrt(trapz(wf[i][3]*wf[i][0]**2, wf[i][0], dx=dx)/norm - aver_x**2)
+    aver_dx = sqrt(trapz(wf[i][3] * wf[i][0] ** 2, wf[i][0], dx=dx) / norm - aver_x ** 2)
     x[2].append(aver_dx)
 x = array(x)