diff --git a/README.md b/README.md
new file mode 100644
index 0000000..2fb307a
--- /dev/null
+++ b/README.md
@@ -0,0 +1,54 @@
+photolib
+========
+
+Cloud-native photo and related media library software, with a focus for photography hobbyists.
+
+TODO longer description
+
+
+Installation
+------------
+
+TODO
+
+
+Usage
+-----
+
+After installation, photolib provides a CLI tool, `photocli`. It can be used as follows:
+
+* `photocli checkdupes` - scan files and print those not already in the library:
+
+
+```
+$ find source -type f -exec shasum -a 256 {} \; | tee shas.txt
+544b5a4f3898abae73260930ee70931a3992c048649692cef89b37576e886f69  source/2018-12-09 13.52.22.jpg
+d2c2c30ca4986d364026644881d667162031008e60aac6a69d7b18230b7ea98c  source/2019-04-08 20.05.02.jpg
+9d42859ed92d7fb978cf73b41293480e1eab03d2d3a14c185c4daf3a49d324ab  source/2019-06-19 16.28.07.png
+...
+
+$ photocli -s http://localhost:8080 checkdupes --sha-files shas.txt
+
+```
+
+TODO: support not using `--sha-files`
+
+
+* `photocli ingest` - import new photos into the library:
+
+
+```
+$ photocli -s http://localhost:8080 ingest photos/*.jpg raws/*.cr2
+
+...
+
+```
+
+TODO: implement this
+
+There is some non-obvious behavior in the ingest process that is important to understand. Many photographers shoot in
+multiple formats, in the sense that one click of the shutter produces multiple formats. Jpeg and Raw is common. During
+ingest, photocli assumes that files with the same filename are multiple formats of the same image. E.g., if you have
+both `IMG_1234.JPG` and `IMG_1234.cr2`, the two images will be grouped in the library. This combination action is only
+applied during ingest; images already in the library that happen to share a filename with a file being imported will
+not be merged. To add a format to a photo already in the library, see `photocli addformat`.
diff --git a/photoapp/cli.py b/photoapp/cli.py
index 4bbe532..c18b066 100644
--- a/photoapp/cli.py
+++ b/photoapp/cli.py
@@ -1,6 +1,8 @@
 import argparse
 import requests
 from requests.exceptions import HTTPError
+from photoapp.utils import get_extension
+from photoapp.types import known_extensions
 
 
 class PhotoApiClient(object):
@@ -32,7 +34,6 @@ def main():
     p_ingest.add_argument("-c", "--copy-of", help="existing uuid the imported images will be placed under")
 
     args = parser.parse_args()
-    print(args)
 
     client = PhotoApiClient(args.host)
 
@@ -48,20 +49,24 @@ def main():
                         if not line:
                             continue
                         sha, path = line.split(maxsplit=1)
-                        hashes[sha] = path.rstrip("\n")
+                        path = path.rstrip("\n")
+                        if get_extension(path) not in known_extensions:
+                            continue
+                        hashes[sha] = path
         else:
             raise NotImplementedError("must pass --sha-files for now")
 
         for sha, path in hashes.items():
             # hit http://localhost:8080/api/v1/byhash?sha=afe49172f709725a4503c9219fb4c6a9db8ad0354fc493f2f500269ac6faeaf6
-            # if the file is a dupe, do nothing
-            # if the file is original, print its path
             try:
-                print(client.byhash(sha))
+                client.byhash(sha)
+                # if the file is a dupe, do nothing
             except HTTPError as he:
-                print(repr(he.response.status_code))
-
-        # print(hashes)
+                # if the file is original, print its path
+                if he.response.status_code == 404:
+                    print(path)
+                else:
+                    raise
 
 
 if __name__ == "__main__":
diff --git a/photoapp/ingest.py b/photoapp/ingest.py
index 4a67c16..28a3e46 100644
--- a/photoapp/ingest.py
+++ b/photoapp/ingest.py
@@ -4,9 +4,10 @@ import traceback
 from photoapp.library import PhotoLibrary
 from photoapp.image import get_jpg_info, get_hash, get_mtime
 from itertools import chain
-from photoapp.types import Photo, PhotoSet
+from photoapp.types import Photo, PhotoSet, known_extensions, regular_images, files_raw, files_video
 import os
 
+
 """
 Photo sorting rules:
 
@@ -20,11 +21,6 @@ mov, video, or other
     modification date
 """
 
-known_extensions = ["jpg", "png", "cr2", "xmp", "mp4", "mov"]
-regular_images = ["jpg", "png"]
-files_raw = ["cr2", "xmp"]
-files_video = ["mp4", "mov"]
-
 
 def pprogress(done, total=None):
     print("  complete: {}{}\r".format(done, " / {} ".format(total) if total else ''), end='')
diff --git a/photoapp/types.py b/photoapp/types.py
index 2e8965e..932c247 100644
--- a/photoapp/types.py
+++ b/photoapp/types.py
@@ -7,6 +7,12 @@ import uuid
 import enum
 
 
+known_extensions = ["jpg", "png", "cr2", "xmp", "mp4", "mov"]
+regular_images = ["jpg", "png"]
+files_raw = ["cr2", "xmp"]
+files_video = ["mp4", "mov"]
+
+
 class PhotoStatus(enum.Enum):
     private = 0
     public = 1
diff --git a/photoapp/utils.py b/photoapp/utils.py
index 9af93f2..74297cd 100644
--- a/photoapp/utils.py
+++ b/photoapp/utils.py
@@ -1,3 +1,4 @@
+import os
 import cherrypy
 from photoapp.types import PhotoSet, PhotoStatus
 
@@ -15,6 +16,13 @@ def mime2ext(mime):
             "video/quicktime": "mov"}[mime]
 
 
+def get_extension(fname):
+    parts = os.path.basename(fname).split(".")
+    if len(parts) == 1:
+        return None
+    return parts[-1].lower()
+
+
 def auth():
     """
     Return the currently authorized username (per request) or None