diff --git a/backend/commands.py b/backend/commands.py
index e88b897..e8a4e02 100644
--- a/backend/commands.py
+++ b/backend/commands.py
@@ -1,15 +1,48 @@
+import libbmc
+import os
+import subprocess
 import tempfile
 
+from backend import config
 from backend import tools
+from libbmc import bibtex
 from libbmc import fetcher
+from libbmc.repositories import arxiv
+from libbmc.papers import identifiers
+from libbmc.papers import tearpages
 
 
-def download(url, filetype, manual, autoconfirm, tag):
+def get_entry_from_index(item, file_or_id=None):
+    """
+    Fetch an entry from the global index.
+
+    :param item: An identifier or filename.
+    :param file_or_id: Whether it is a file or an entry identifier. If \
+            ``None``, will try to match both.
+    :returns: TODO.
+    """
+    entry = None
+    # If explictly an identifier
+    if file_or_id == "id":
+        entry = bibtex.get_entry(config.get("index"), item)
+    # If explicitely a filename
+    elif file_or_id == "file":
+        entry = bibtex.get_entry_by_filter(config.get("index"),
+                                           lambda x: x.file == item)  # TODO
+    # Else, get by id or file
+    else:
+        entry = bibtex.get_entry_by_filter(config.get("index"),
+                                           lambda x: (
+                                               x.id == item or
+                                               x.file == item))  # TODO
+    return entry
+
+
+def download(url, manual, autoconfirm, tag):
     """
     Download a given URL and add it to the library.
 
     :param url: URL to download.
-    :param filetype: paper / book / ``None``.
     :param manual: Whether BibTeX should be fetched automatically.
     :param autoconfirm: Whether import should be made silent or not.
     :param tag: A tag for this file.
@@ -30,7 +63,7 @@ def download(url, filetype, manual, autoconfirm, tag):
                 fh.write(dl)
 
             # And add it as a normal paper from now on
-            new_name = import_file(tmp.name, filetype, manual,
+            new_name = import_file(tmp.name, manual,
                                    autoconfirm, tag)
             if new_name is None:
                 return None
@@ -43,17 +76,200 @@ def download(url, filetype, manual, autoconfirm, tag):
         return None
 
 
-def import_file(src, filetype, manual, autoconfirm, tag, rename=True):
+def import_file(src, manual=False, autoconfirm=False,
+                tag='', rename=True):
     """
     Add a file to the library.
 
     :param src: The path of the file to import.
-    :param filetype: paper / book / ``None``.
-    :param manual: Whether BibTeX should be fetched automatically.
-    :param autoconfirm: Whether import should be made silent or not.
-    :param tag: A tag for this file.
-    :param rename: TODO
+    :param manual: Whether BibTeX should be fetched automatically. \
+            Default to ``False``.
+    :param autoconfirm: Whether import should be made silent or not. \
+            Default to ``False``.
+    :param tag: A tag for this file. \
+            Default to no tag.
+    :param rename: Whether or not the file should be renamed according to the \
+            mask in the config.
     :returns: The name of the imported file, or ``None`` in case of error.
     """
+    if not manual:
+        type, identifier = identifiers.find_identifiers(src)
+
+    if type is None:
+        tools.warning("Could not find an identifier for %s. \
+                      Switching to manual entry." % (src))
+        # Fetch available identifiers types from libbmc
+        # Append "manual" for manual entry of BibTeX and "skip" to skip the
+        # file.
+        available_types_list = (libbmc.__valid_identifiers__ +
+                                ["manual", "skip"])
+        available_types = " / ".joint(available_types_list)
+        # Query for the type to use
+        while type not in available_types_list:
+            type = input("%s? " % (available_types)).lower()
+
+        if type == "skip":
+            # If "skip" is chosen, skip this file
+            return None
+        elif type == "manual":
+            identifier = None
+        else:
+            # Query for the identifier if required
+            identifier = input("Value? ")
+    else:
+        print("%s found for %s: %s." % (type, src, identifier))
+
+    # Fetch BibTeX automatically if we have an identifier
+    bibtex = None
+    if identifier is not None:
+        # If an identifier was provided, try to automatically fetch the bibtex
+        bibtex = identifiers.get_bibtex((type, identifier))
+
+    # TODO: Check bibtex
+
+    # Handle tag
+    if not autoconfirm:
+        # If autoconfirm is not enabled, query for a tag
+        user_tag = input("Tag for this paper [%s]? " % tag)
+        if user_tag != "":
+            tag = user_tag
+    bibtex["tag"] = tag
+
+    # TODO: Handle renaming
+    new_name = src
+    if rename:
+        pass
+
+    bibtex['file'] = os.path.abspath(new_name)
+
+    # Tear some pages if needed
+    should_tear_pages = True
+    if not autoconfirm:
+        # Ask for confirmation
+        pages_to_tear = tearpages.tearpage_needed(bibtex)
+        user_tear_pages = input("Found some pages to tear: %s. \
+                                Confirm? [Y/n]" % (pages_to_tear)).lower()
+        if user_tear_pages == "n":
+            should_tear_pages = False
+    if should_tear_pages:
+        tearpages.tearpage(new_name, bibtex=bibtex)
+
+    # TODO: Append to global bibtex index
+
+    return new_name
+
+
+def delete(item, keep=False, file_or_id=None):
+    """
+    Delete an entry in the main BibTeX file, and the associated documents.
+
+    :param item: An entry or filename to delete from the database.
+    :param keep: Whether or not the document should be kept on the disk. \
+            If True, will simply delete the entry in the main BibTeX index.
+    :param file_or_id: Whether it is a file or an entry identifier. If \
+            ``None``, will try to match both.
+    :returns: Nothing.
+    """
+    entry = get_entry_from_index(item, file_or_id)
+
+    # Delete the entry from the bibtex index
+    bibtex.delete(config.get("index"), entry.id)  # TODO
+
+    # If file should not be kept
+    if not keep:
+        # Delete it
+        os.unlink(entry.file)  # TODO
+
+
+def edit(item, file_or_id):
+    """
+    Edit an entry in the main BibTeX file.
+
+    :param item: An entry or filename to edit in the database.
+    :param file_or_id: Whether it is a file or an entry identifier. If \
+            ``None``, will try to match both.
+    :returns: Nothing.
+    """
     # TODO
     pass
+
+
+def list_entries():
+    """
+    List all the available entries and their associated files.
+
+    :returns: A dict with entry identifiers as keys and associated files as \
+            values.
+    """
+    # Get the list of entries from the BibTeX index
+    entries_list = bibtex.get(config.get("index"))
+    return {entry.id: entry.file for entry in entries_list}  # TODO
+
+
+def open(id):
+    """
+    Open the file associated with the provided entry identifier.
+
+    :param id: An entry identifier in the main BibTeX file.
+    :returns: ``False`` if an error occured. ``True`` otherwise.
+    """
+    # Fetch the entry from the BibTeX index
+    entry = bibtex.get_entry(config.get("index"), id)
+    if entry is None:
+        return False
+    else:
+        # Run xdg-open on the associated file to open it
+        subprocess.Popen(['xdg-open', entry.filename])  # TODO
+        return True
+
+
+def export(item, file_or_id=None):
+    """
+    Export the BibTeX entries associated to some items.
+
+    :param item: An entry or filename to export as BibTeX.
+    :param file_or_id: Whether it is a file or an entry identifier. If \
+            ``None``, will try to match both.
+    :returns: TODO.
+    """
+    # Fetch the entry from the BibTeX index
+    entry = get_entry_from_index(item, file_or_id)
+    if entry is not None:
+        return bibtex.dict2BibTeX(entry)  # TODO
+
+
+def resync():
+    """
+    Compute the diff between the main BibTeX index and the files on the disk,
+    and try to resync them.
+
+    :returns: Nothing.
+    """
+    # TODO
+    pass
+
+
+def update(item, file_or_id=None):
+    """
+    Update an entry, trying to fetch a more recent version (on arXiv for \
+            instance.)
+
+    :param item: An entry or filename to fetch update from.
+    :param file_or_id: Whether it is a file or an entry identifier. If \
+            ``None``, will try to match both.
+    :returns: TODO.
+    """
+    entry = get_entry_from_index(item, file_or_id)
+    # Fetch latest version
+    latest_version = arxiv.get_latest_version(entry.eprint)  # TODO
+    if latest_version != entry.eprint:  # TODO
+        print("New version found for %s: %s" % (entry, latest_version))
+        confirm = input("Download it? [Y/n] ")
+        if confirm.lower() == 'n':
+            return
+
+        # Download the updated version
+        # TODO
+
+        # Delete previous version if needed
+        # TODO
diff --git a/backend/config.py b/backend/config.py
index 3c78f6c..f28329a 100644
--- a/backend/config.py
+++ b/backend/config.py
@@ -55,6 +55,7 @@ class Config():
 
     def initialize(self):
         self.set("folder", os.path.expanduser("~/Papers/"))
+        self.set("index", os.path.expanduser("~/Papers/index.bib"))
         self.set("proxies", [''])
         self.set("format_articles", "%f_%l-%j-%Y%v")
         self.set("format_books", "%a-%t")
diff --git a/bmc.py b/bmc.py
index e3836b6..d996377 100755
--- a/bmc.py
+++ b/bmc.py
@@ -15,6 +15,14 @@ from backend.config import Config
 EDITOR = os.environ.get("EDITOR")
 
 
+def file_or_id_from_args(args):
+    """
+    Helper function to parse provided args to check if the argument is a \
+            file or an identifier.
+    """
+    return "id" if args.id else "file" if args.file else None
+
+
 def parse_args():
     """
     Build a parser and parse arguments of command line.
@@ -25,13 +33,10 @@ def parse_args():
         description="A bibliography management tool.")
     parser.add_argument("-c", "--config", default=None,
                         help="path to a custom config dir.")
-    subparsers = parser.add_subparsers(help="sub-command help", dest='parser')
+    subparsers = parser.add_subparsers(help="sub-command help", dest='command')
     subparsers.required = True  # Fix for Python 3.3.5
 
     parser_download = subparsers.add_parser('download', help="download help")
-    parser_download.add_argument('-t', '--type', default=None,
-                                 choices=['paper', 'book'],
-                                 help="type of the file to download")
     parser_download.add_argument('-m', '--manual', default=False,
                                  action='store_true',
                                  help="disable auto-download of bibtex")
@@ -46,9 +51,6 @@ def parse_args():
     parser_download.set_defaults(func='download')
 
     parser_import = subparsers.add_parser('import', help="import help")
-    parser_import.add_argument('-t', '--type', default=None,
-                               choices=['paper', 'book'],
-                               help="type of the file to import")
     parser_import.add_argument('-m', '--manual', default=False,
                                action='store_true',
                                help="disable auto-download of bibtex")
@@ -100,16 +102,30 @@ def parse_args():
     parser_open.set_defaults(func='open')
 
     parser_export = subparsers.add_parser('export', help="export help")
-    parser_export.add_argument('ids', metavar='id',  nargs='+',
-                               help="an identifier")
+    parser_export.add_argument('entries', metavar='entry', nargs='+',
+                               help="a filename or an identifier")
+    parser_export.add_argument('--skip',  nargs='+',
+                               help="path to files to skip", default=[])
+    group = parser_export.add_mutually_exclusive_group()
+    group.add_argument('--id', action="store_true", default=False,
+                       help="id based deletion")
+    group.add_argument('--file', action="store_true", default=False,
+                       help="file based deletion")
     parser_export.set_defaults(func='export')
 
     parser_resync = subparsers.add_parser('resync', help="resync help")
     parser_resync.set_defaults(func='resync')
 
     parser_update = subparsers.add_parser('update', help="update help")
-    parser_update.add_argument('--entries', metavar='entry', nargs='+',
+    parser_update.add_argument('entries', metavar='entry', nargs='+',
                                help="a filename or an identifier")
+    parser_update.add_argument('--skip',  nargs='+',
+                               help="path to files to skip", default=[])
+    group = parser_update.add_mutually_exclusive_group()
+    group.add_argument('--id', action="store_true", default=False,
+                       help="id based deletion")
+    group.add_argument('--file', action="store_true", default=False,
+                       help="file based deletion")
     parser_update.set_defaults(func='update')
 
     return parser.parse_args()
@@ -133,7 +149,7 @@ def main():
         skipped = []
         for url in args.url:
             # Try to download the URL
-            new_name = commands.download(url, args.type, args.manual, args.y,
+            new_name = commands.download(url, args.manual, args.y,
                                          args.tag)
             if new_name is not None:
                 print("%s successfully imported as %s." % (url, new_name))
@@ -154,7 +170,7 @@ def main():
         files_to_process = list(set(args.file) - set(args.skip))
         for filename in files_to_process:
             # Try to import the file
-            new_name = commands.import_file(filename, args.type,
+            new_name = commands.import_file(filename,
                                             args.manual, args.y,
                                             args.tag, not args.inplace)
             if new_name is not None:
@@ -184,7 +200,7 @@ def main():
 
             # Try to delete the item
             if confirm.lower() == 'y':
-                file_or_id = "id" if args.id else "file" if args.file else None
+                file_or_id = file_or_id_from_args(args)
                 commands.delete(item, args.keep, file_or_id)
                 print("%s successfully deleted." % (item,))
             else:
@@ -200,13 +216,13 @@ def main():
         # Handle exclusions
         items_to_process = list(set(args.entries) - set(args.skip))
         for item in items_to_process:
-            file_or_id = "id" if args.id else "file" if args.file else None
+            file_or_id = file_or_id_from_args(args)
             commands.edit(item, file_or_id)
 
     # List command
     elif args.func == 'list':
         # List all available items
-        for id, file in commands.list().items():
+        for id, file in commands.list_entries().items():
             # And print them as "identifier: file"
             print("%s: %s" % (id, file))
 
@@ -214,14 +230,18 @@ def main():
     elif args.func == 'open':
         # Open each entry
         for id in args.ids:
-            if commands.open(id) is None:
+            if not commands.open(id):
                 # And warn the user about missing files or errors
                 tools.warning("Unable to open file associated with ident %s." %
                               (id,))
 
     # Export command
     elif args.func == 'export':
-        print(commands.export(args.ids))
+        # Handle exclusions
+        items_to_process = list(set(args.entries) - set(args.skip))
+        for item in items_to_process:
+            file_or_id = file_or_id_from_args(args)
+            print(commands.export(item, file_or_id))
 
     # Resync command
     elif args.func == 'resync':
@@ -231,7 +251,15 @@ def main():
 
     # Update command
     elif args.func == 'update':
-        commands.update(args.entries)
+        # Handle exclusions
+        items_to_process = list(set(args.entries) - set(args.skip))
+        for item in items_to_process:
+            file_or_id = file_or_id_from_args(args)
+            updates = commands.update(args.entries)
+            # TODO \/
+            print("%d new versions of papers were found:" % (len(updates)))
+            for item in updates:
+                print(item)
 
 
 if __name__ == '__main__':