diff --git a/README.md b/README.md
index 992bf7d..77a7b37 100644
--- a/README.md
+++ b/README.md
@@ -178,6 +178,39 @@ Y:\DATA\SARDANA\MIHC\768473\RAW
             768473.ome.tif.log
 ```
 
+### Align whole-slide multiplexed IF images
+
+```bash
+palom-multicycle-tif --img_list '/path/cycle1.tif' '/path/to/cycle2.tif' '/path/to/cycle3.tif' --out_dir '/path/to/outdir' --out_name 'example.ome.tif'
+```
+
+Parameters:
+`--img_list`: provide a list of whole-slide images to be registered in `tif` or `tiff` format
+`--out_dir`: provide a path to the output directory
+`--out_name`: optional, provide output file name. If not provided, `-registered.ome.tif` will be appended to the stem of the second image path
+`--channel`: optional, channel to guide alignment. If not provided, default value 0
+`--px_size`: optional, physical pixel size of the image. The image metadata will be checked if not provided. If not found, value will be set to 1. 
+`--n_keypoints`: optional, number of keypoints to use for alignment. Default 10000.
+`--only_coarse`: bool, only run coarse alignment
+`--only_qc`: bool, only run quality control
+
+When the process is finished, a pyramidal OME-TIFF file will be generated along
+with PNG files showing the feature-based registration results.
+
+```
+/path/to/
+│   cycle1.tif
+│   cycle2.tif
+│   cycle3.tif
+│
+└───outdir
+    │   example.ome.tif
+    │
+    └───qc
+            '1-cycle2.ome.tif (block shift distance).png'
+            '1-cycle2.ome.tif (coarse alignment).jpg'
+```
+
 ---
 
 ## Scripting
diff --git a/palom/cli/align_multiple_cycles.py b/palom/cli/align_multiple_cycles.py
new file mode 100644
index 0000000..7ca6945
--- /dev/null
+++ b/palom/cli/align_multiple_cycles.py
@@ -0,0 +1,230 @@
+import pathlib
+import argparse
+import matplotlib
+import matplotlib.pyplot as plt
+import palom
+import sys
+
+def parse_args():
+    parser = argparse.ArgumentParser(
+        description=(
+            'Align multiple cycles of images using PALOM'
+        )   
+    )
+    parser.add_argument(
+        '--img_list',
+        dest='img_list',
+        nargs='+',
+        help='list of image paths to be aligned'
+    )
+    parser.add_argument(
+        '--out_dir',
+        dest='out_dir',
+        help='output directory'
+    )
+    parser.add_argument(
+        '--out_name',
+        dest='out_name',
+        help='output file name'
+    )
+    parser.add_argument(
+        '--thumbnail_level',
+        metavar='thumbnail_level',
+        type=int,
+        default=None,
+        help='Level to use for coarse alignment'
+    )
+    parser.add_argument(
+        '--channel',
+        dest='channel',
+        type=int,
+        default=0,
+        help='channel to use for alignment'
+    )
+    parser.add_argument(
+        '--px_size',
+        dest='px_size',
+        type=float,
+        help='pixel size'
+    )
+    parser.add_argument(
+        '--n_keypoints',
+        dest='n_keypoints',
+        type=int,
+        default=10000,
+        help='number of keypoints'
+    )
+    parser.add_argument(
+        '--auto_mask',
+        dest='auto_mask',
+        type=bool,
+        default=True,
+        help='auto mask'
+    )
+    parser.add_argument(
+        '--only_coarse',
+        dest='only_coarse',
+        action='store_true',
+        help='only coarse alignment'
+    )
+    parser.add_argument(
+        '--only_qc',
+        dest='only_qc',
+        type=bool,
+        default=False,
+        help='only quality control'
+    )
+    arg = parser.parse_args()
+
+    return arg
+
+def align_multiple_cycles(args):
+    #thumbnail_channel: int = 1,
+
+    only_coarse = args.only_coarse
+    only_qc = args.only_qc
+    auto_mask = args.auto_mask
+    n_keypoints = args.n_keypoints
+    px_size = args.px_size
+
+    out_dir = pathlib.Path(args.out_dir)
+    img_list = [pathlib.Path(img_path) for img_path in args.img_list]
+    assert len(img_list) >= 2, "At least two images are required for registration"
+
+    if args.out_name is None:
+        out_name = f"{img_list[1].stem}-registered.ome.tif"
+    else:
+        out_name = args.out_name
+    channel = args.channel
+
+    out_path = out_dir / out_name
+    assert ''.join(out_path.suffixes[-2:]) in ('.ome.tif', '.ome.tiff')
+    out_path.parent.mkdir(exist_ok=True, parents=True)
+    set_matplotlib_font(font_size=8)
+
+    ref_reader = palom.reader.OmePyramidReader(img_list[0], pixel_size=px_size)
+    moving_readers = [palom.reader.OmePyramidReader(file) for file in img_list[1:]]
+
+    LEVEL = 0
+    if args.thumbnail_level is None:
+        THUMBNAIL_LEVEL = ref_reader.get_thumbnail_level_of_size(2000)
+    else:
+        THUMBNAIL_LEVEL = args.thumbnail_level
+
+    mosaic_list = [ref_reader.pyramid[LEVEL]]
+    for idx, moving_reader in enumerate(moving_readers):
+        aligner = palom.align.Aligner(
+                # use the first channel (Hoechst staining) in the reference image as the
+                # registration reference
+                ref_img=ref_reader.read_level_channels(LEVEL, channel),
+                # use the second channel (G channel) in the moving image, it usually has
+                # better contrast
+                moving_img=moving_reader.read_level_channels(LEVEL, channel),
+                # select the same channels for the thumbnail images
+                ref_thumbnail=ref_reader.read_level_channels(THUMBNAIL_LEVEL, channel).compute(),
+                moving_thumbnail=moving_reader.read_level_channels(THUMBNAIL_LEVEL, channel).compute(),
+                # specify the downsizing factors so that the affine matrix can be scaled to
+                # match the registration reference
+                ref_thumbnail_down_factor=ref_reader.level_downsamples[THUMBNAIL_LEVEL] /ref_reader.level_downsamples[LEVEL],
+                moving_thumbnail_down_factor=moving_reader.level_downsamples[THUMBNAIL_LEVEL] / moving_reader.level_downsamples[LEVEL]
+            )
+
+        aligner.coarse_register_affine(
+            n_keypoints=n_keypoints,
+            test_flip=True,
+            test_intensity_invert=True,
+            auto_mask=auto_mask
+        )
+
+        fig, ax = plt.gcf(), plt.gca()
+        fig.suptitle(f"{img_list[idx+1].name} (coarse alignment)", fontsize=8)
+        ax.set_title(f"{img_list[0].name} - {img_list[idx+1].name}", fontsize=6)
+        im_h, im_w = ax.images[0].get_array().shape
+        set_subplot_size(im_w/288, im_h/288, ax=ax)
+        ax.set_anchor('N')
+        # use 0.5 inch on the top for figure title
+        fig.subplots_adjust(top=1 - .5 / fig.get_size_inches()[1])
+        save_all_figs(out_dir=out_dir / 'qc', format='jpg', dpi=144)
+
+        if not only_coarse:
+            
+            aligner.compute_shifts()
+        
+            fig = aligner.plot_shifts()
+            fig.suptitle(f"{img_list[idx+1].name} (block shift distance)", fontsize=8)
+            fig.axes[0].set_title(img_list[0].name, fontsize=6)
+            save_all_figs(out_dir=out_dir / 'qc', format='png')
+
+            aligner.constrain_shifts()
+            block_mx = aligner.block_affine_matrices_da
+
+   
+        if not only_qc:
+            mx = aligner.affine_matrix
+            if not only_coarse:
+                mx = block_mx
+            
+            mosaic = palom.align.block_affine_transformed_moving_img(
+                ref_img=ref_reader.read_level_channels(LEVEL, channel),
+                moving_img=moving_reader.pyramid[LEVEL],
+                mxs=mx
+            )
+            mosaic_list.append(mosaic)
+
+    if not only_coarse and not only_qc:
+
+        palom.pyramid.write_pyramid(
+            mosaics=mosaic_list,
+            output_path=out_path,
+            pixel_size=ref_reader.pixel_size*ref_reader.level_downsamples[LEVEL],
+            compression='zlib',
+            downscale_factor=2,
+            save_RAM=True,
+            tile_size=1024
+        )
+    return 0
+
+
+def set_matplotlib_font(font_size=12):
+    font_families = matplotlib.rcParams['font.sans-serif']
+    if font_families[0] != 'Arial':
+        font_families.insert(0, 'Arial')
+    matplotlib.rcParams['pdf.fonttype'] = 42
+    matplotlib.rcParams.update({'font.size': font_size})
+
+
+def save_all_figs(dpi=300, format='pdf', out_dir=None, prefix=None):
+    figs = [plt.figure(i) for i in plt.get_fignums()]
+    if prefix is not None:
+        for f in figs:
+            if f._suptitle:
+                f.suptitle(f"{prefix} {f._suptitle.get_text()}")
+            else:
+                f.suptitle(prefix)
+    names = [f._suptitle.get_text() if f._suptitle else '' for f in figs]
+    out_dir = pathlib.Path(out_dir)
+    out_dir.mkdir(exist_ok=True, parents=True)
+
+    for f, n, nm in zip(figs, plt.get_fignums(), names):
+        f.savefig(out_dir / f"{n}-{nm}.{format}", dpi=dpi, bbox_inches='tight')
+        plt.close(f)
+
+
+def set_subplot_size(w, h, ax=None):
+    """ w, h: width, height in inches """
+    if not ax: ax=plt.gca()
+    l = ax.figure.subplotpars.left
+    r = ax.figure.subplotpars.right
+    t = ax.figure.subplotpars.top
+    b = ax.figure.subplotpars.bottom
+    figw = float(w)/(r-l)
+    figh = float(h)/(t-b)
+    ax.figure.set_size_inches(figw, figh)
+
+def main():
+    align_multiple_cycles(parse_args())
+
+if __name__ == '__main__':
+    sys.exit(main())
+
+    
\ No newline at end of file
diff --git a/pyproject.toml b/pyproject.toml
index 5cff201..aa22274 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -43,6 +43,7 @@ build-backend = "poetry.core.masonry.api"
 [tool.poetry.scripts]
 palom-svs = "palom.cli.svs:main"
 palom-svs-helper = "palom.cli.helper:main"
+palom-multicycle-tif = "palom.cli.align_multiple_cycles:main"
 palom-align-he = "palom.cli.align_he:main"
 palom-pyramid = "palom.cli.pyramid_tools:main"
-palom-rgb-compress = "palom.cli.compress_rgb_jpeg:main"
\ No newline at end of file
+palom-rgb-compress = "palom.cli.compress_rgb_jpeg:main"