diff --git a/invokeai/app/util/read_watermark.py b/invokeai/app/util/read_watermark.py
new file mode 100644
index 00000000000..5fe492d2b04
--- /dev/null
+++ b/invokeai/app/util/read_watermark.py
@@ -0,0 +1,34 @@
+"""CLI command to decode invisible watermarks from Invoke-generated images."""
+
+import argparse
+import sys
+
+from PIL import Image
+
+from invokeai.backend.image_util.invisible_watermark import InvisibleWatermark
+
+
+def read_watermark() -> None:
+    """Read and print invisible watermarks from a list of image files."""
+    parser = argparse.ArgumentParser(
+        prog="invoke-readwatermark",
+        description="Decode invisible watermarks from Invoke-generated images.",
+    )
+    parser.add_argument("images", nargs="+", metavar="IMAGE", help="Image file(s) to read watermarks from.")
+    parser.add_argument(
+        "--length",
+        type=int,
+        default=8,
+        metavar="BYTES",
+        help="Expected watermark length in bytes (default: %(default)s, matching the default 'InvokeAI' watermark text).",
+    )
+    args = parser.parse_args()
+
+    for path in args.images:
+        try:
+            image = Image.open(path)
+        except OSError as e:
+            print(f"{path}: error opening image: {e}", file=sys.stderr)
+            continue
+        watermark = InvisibleWatermark.decode_watermark(image, watermark_length=args.length)
+        print(f"{path}: {watermark}")
diff --git a/invokeai/backend/image_util/invisible_watermark.py b/invokeai/backend/image_util/invisible_watermark.py
index 5b0b2dbb5b1..6a87f4da47e 100644
--- a/invokeai/backend/image_util/invisible_watermark.py
+++ b/invokeai/backend/image_util/invisible_watermark.py
@@ -9,7 +9,7 @@
 from PIL import Image
 
 import invokeai.backend.util.logging as logger
-from invokeai.backend.image_util.imwatermark.vendor import WatermarkEncoder
+from invokeai.backend.image_util.imwatermark.vendor import WatermarkDecoder, WatermarkEncoder
 
 
 class InvisibleWatermark:
@@ -25,3 +25,23 @@ def add_watermark(cls, image: Image.Image, watermark_text: str) -> Image.Image:
         encoder.set_watermark("bytes", watermark_text.encode("utf-8"))
         bgr_encoded = encoder.encode(bgr, "dwtDct")
         return Image.fromarray(cv2.cvtColor(bgr_encoded, cv2.COLOR_BGR2RGB)).convert("RGBA")
+
+    @classmethod
+    def decode_watermark(cls, image: Image.Image, watermark_length: int = 8) -> str:
+        """Decode an invisible watermark from an image.
+
+        Args:
+            image: The PIL image to decode the watermark from.
+            watermark_length: The length of the watermark in bytes (default 8, matching the default "InvokeAI" watermark text).
+
+        Returns:
+            The decoded watermark text, or an empty string if decoding fails.
+        """
+        bgr = cv2.cvtColor(np.array(image.convert("RGB")), cv2.COLOR_RGB2BGR)
+        decoder = WatermarkDecoder("bytes", watermark_length * 8)
+        try:
+            raw = decoder.decode(bgr, "dwtDct")
+            return raw.rstrip(b"\x00").decode("utf-8", errors="replace")
+        except (RuntimeError, ValueError, NameError) as e:
+            logger.debug("Failed to decode watermark: %s", e)
+            return ""
diff --git a/pyproject.toml b/pyproject.toml
index adfe5982baf..a7cf0d3a420 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -159,6 +159,7 @@ explicit = true
 
 [project.scripts]
 "invokeai-web" = "invokeai.app.run_app:run_app"
+"invoke-readwatermark" = "invokeai.app.util.read_watermark:read_watermark"
 
 [project.urls]
 "Homepage" = "https://invoke-ai.github.io/InvokeAI/"