diff --git a/src/borg/archive.py b/src/borg/archive.py
index 16ab75e31e..3ea0f413cb 100644
--- a/src/borg/archive.py
+++ b/src/borg/archive.py
@@ -1,4 +1,5 @@
 import errno
+import io
 import os
 import socket
 import stat
@@ -22,7 +23,7 @@
 from .compress import Compressor
 from .constants import *  # NOQA
 from .hashindex import ChunkIndex, ChunkIndexEntry
-from .helpers import Manifest
+from .helpers import Manifest, slice_chunks
 from .helpers import Chunk, ChunkIteratorFileWrapper, open_item
 from .helpers import Error, IntegrityError
 from .helpers import uid2user, user2uid, gid2group, group2gid
@@ -428,7 +429,7 @@ def add_file_chunks(chunks):
         return stats
 
     def extract_item(self, item, restore_attrs=True, dry_run=False, stdout=False, sparse=False,
-                     hardlink_masters=None, stripped_components=0, original_path=None, pi=None):
+                     hardlink_masters=None, complete_partial=False, stripped_components=0, original_path=None, pi=None):
         """
         Extract archive item.
 
@@ -437,6 +438,7 @@ def extract_item(self, item, restore_attrs=True, dry_run=False, stdout=False, sp
         :param dry_run: do not write any data
         :param stdout: write extracted data to stdout
         :param sparse: write sparse files (chunk-granularity, independent of the original being sparse)
+        :param complete_partial: False: replace files, True: existing files are completed
         :param hardlink_masters: maps paths to (chunks, link_target) for extracting subtrees with hardlinks correctly
         :param stripped_components: stripped leading path components to correct hard link extraction
         :param original_path: 'path' key as stored in archive
@@ -466,14 +468,15 @@ def extract_item(self, item, restore_attrs=True, dry_run=False, stdout=False, sp
         # Attempt to remove existing files, ignore errors on failure
         try:
             st = os.lstat(path)
-            if stat.S_ISDIR(st.st_mode):
-                os.rmdir(path)
-            else:
-                os.unlink(path)
+            if not complete_partial:
+                if stat.S_ISDIR(st.st_mode):
+                    os.rmdir(path)
+                else:
+                    os.unlink(path)
         except UnicodeEncodeError:
             raise self.IncompatibleFilesystemEncodingError(path, sys.getfilesystemencoding()) from None
         except OSError:
-            pass
+            st = None
         mode = item.mode
         if stat.S_ISREG(mode):
             with backup_io():
@@ -496,9 +499,30 @@ def extract_item(self, item, restore_attrs=True, dry_run=False, stdout=False, sp
                     return
                 # Extract chunks, since the item which had the chunks was not extracted
             with backup_io():
-                fd = open(path, 'wb')
+                if complete_partial and st is not None:
+                    # Open existing file for updating
+                    # Note that 'ab' wouldn't work on e.g. NetBSD, since seeking would be meaningless for writes.
+                    # However, r+b requires the file to exist. Therefore we need to distinguish the two cases here.
+                    fd = open(path, 'r+b')
+                else:
+                    fd = open(path, 'wb')
             with fd:
-                ids = [c.id for c in item.chunks]
+                chunks = item.chunks
+                if complete_partial:
+                    with backup_io():
+                        fd.seek(0, io.SEEK_END)
+                        existing_length = fd.tell()
+                        # Slice chunks by current length of the existing file.
+                        chunks, prefix_length = slice_chunks(chunks, maximum_length=existing_length)
+                        # We don't bother extracting fractional chunks. Just seek to a chunk boundary.
+                        fd.seek(prefix_length)
+                        fd.truncate()
+                    discarded_count = len(item.chunks) - len(chunks)
+                    discarded_chunks_ids = [c.id for c in item.chunks[:discarded_count]]
+                    self.repository.discard_preload(discarded_chunks_ids)
+                    if pi:
+                        pi.show(increase=prefix_length)
+                ids = [c.id for c in chunks]
                 for _, data in self.pipeline.fetch_many(ids, is_preloaded=True):
                     if pi:
                         pi.show(increase=len(data))
diff --git a/src/borg/archiver.py b/src/borg/archiver.py
index 3abe7a5069..790a91b38a 100644
--- a/src/borg/archiver.py
+++ b/src/borg/archiver.py
@@ -487,11 +487,16 @@ def do_extract(self, args, repository, manifest, key, archive):
         dry_run = args.dry_run
         stdout = args.stdout
         sparse = args.sparse
+        continue_ = args.continue_
         strip_components = args.strip_components
         dirs = []
         partial_extract = not matcher.empty() or strip_components
         hardlink_masters = {} if partial_extract else None
 
+        if stdout and continue_:
+            self.print_error('Cannot combine --stdout and --continue.')
+            return self.exit_code
+
         def peek_and_store_hardlink_masters(item, matched):
             if (partial_extract and not matched and stat.S_ISREG(item.mode) and
                     item.get('hardlink_master', True) and 'source' not in item):
@@ -527,8 +532,9 @@ def peek_and_store_hardlink_masters(item, matched):
                         dirs.append(item)
                         archive.extract_item(item, restore_attrs=False)
                     else:
-                        archive.extract_item(item, stdout=stdout, sparse=sparse, hardlink_masters=hardlink_masters,
-                                             stripped_components=strip_components, original_path=orig_path, pi=pi)
+                        archive.extract_item(item, stdout=stdout, complete_partial=continue_, sparse=sparse,
+                                             hardlink_masters=hardlink_masters, original_path=orig_path, pi=pi,
+                                             stripped_components=strip_components)
             except BackupOSError as e:
                 self.print_warning('%s: %s', remove_surrogates(orig_path), e)
 
@@ -1867,6 +1873,9 @@ def build_parser(self, prog=None):
         subparser.add_argument('--sparse', dest='sparse',
                                action='store_true', default=False,
                                help='create holes in output sparse file from all-zero chunks')
+        subparser.add_argument('--continue', dest='continue_',
+                               action='store_true', default=False,
+                               help='continue interrupted extraction')
         subparser.add_argument('location', metavar='ARCHIVE',
                                type=location_validator(archive=True),
                                help='archive to extract')
diff --git a/src/borg/helpers.py b/src/borg/helpers.py
index 6d6b8c7e75..bb7fe70846 100644
--- a/src/borg/helpers.py
+++ b/src/borg/helpers.py
@@ -23,7 +23,7 @@
 from datetime import datetime, timezone, timedelta
 from fnmatch import translate
 from functools import wraps, partial, lru_cache
-from itertools import islice
+from itertools import islice, dropwhile
 from operator import attrgetter
 from string import Formatter
 
@@ -1769,3 +1769,21 @@ def swidth_slice(string, max_width):
     if reverse:
         result.reverse()
     return ''.join(result)
+
+
+def slice_chunks(chunks, maximum_length):
+    """
+    Slice *chunks* (list(ChunkListEntry)) to remove a prefix of *maximum_length*.
+
+    Return (sliced_chunks, prefix_length).
+    """
+    def should_drop(chunk):
+        nonlocal current_length
+        dropped = (current_length + chunk.size) <= maximum_length
+        if dropped:
+            current_length += chunk.size
+        return dropped
+
+    current_length = 0
+    sliced_chunks = dropwhile(should_drop, chunks)
+    return list(sliced_chunks), current_length
diff --git a/src/borg/remote.py b/src/borg/remote.py
index 294bd40cf7..e029bc6ef2 100644
--- a/src/borg/remote.py
+++ b/src/borg/remote.py
@@ -506,6 +506,22 @@ def close(self):
     def preload(self, ids):
         self.preload_ids += ids
 
+    def discard_preload(self, discard_ids):
+        # Stop preloading these immediately
+        self.preload_ids[:] = [id for id in self.preload_ids
+                               if id not in discard_ids]
+
+        for discard_id in discard_ids:
+            msgids = self.cache.get((discard_id,), [])
+            for msgid in list(msgids):
+                if msgid in self.responses:
+                    # Discard anything we may have already received
+                    self.responses.pop(msgid)
+                    msgids.remove(msgid)
+                else:
+                    # The call was sent out, but no response yet; if we get one, we'll ignore it
+                    self.ignore_responses.add(msgid)
+
 
 def handle_remote_line(line):
     if line.startswith('$LOG '):
diff --git a/src/borg/repository.py b/src/borg/repository.py
index 095a9ba7f5..a98c60b1dd 100644
--- a/src/borg/repository.py
+++ b/src/borg/repository.py
@@ -805,6 +805,14 @@ def preload(self, ids):
         """Preload objects (only applies to remote repositories)
         """
 
+    def discard_preload(self, discard_ids):
+        """
+        Ignore/stop preloading objects (only applies to remote repositories).
+
+        Be warned: trying to get_many() anything from discard_ids while specifying is_preloaded=True *will*
+        deadlock with remote repositories.
+        """
+
 
 class LoggedIO:
 
diff --git a/src/borg/testsuite/archiver.py b/src/borg/testsuite/archiver.py
index 50cb42c5c4..b2fa810e72 100644
--- a/src/borg/testsuite/archiver.py
+++ b/src/borg/testsuite/archiver.py
@@ -736,6 +736,32 @@ def test_extract_with_pattern(self):
             self.cmd("extract", self.repository_location + "::test", "fm:input/file1", "fm:*file33*", "input/file2")
         self.assert_equal(sorted(os.listdir("output/input")), ["file1", "file2", "file333"])
 
+    def test_extract_continue(self):
+        self.cmd('init', self.repository_location)
+        self.create_regular_file('file1', size=1024 * 80)
+        self.create_regular_file('file2', size=1024 * 100)
+        self.cmd('create', self.repository_location + '::test', 'input')
+
+        with changedir('output'):
+            self.cmd('extract', self.repository_location + '::test')
+            with open('input/file1', 'ab') as file:
+                file.truncate(1234)
+            with open('input/file2', 'ab') as file:
+                file.write(b'extra bytes')
+            self.cmd('extract', self.repository_location + '::test', '--continue')
+        self.assert_dirs_equal('input', 'output/input')
+
+        shutil.rmtree('output/input')
+
+        with changedir('output'):
+            os.mkdir('input')
+            with open('input/file1', 'ab') as file:
+                file.truncate(1234)
+            with open('input/file2', 'ab') as file:
+                file.write(b'extra bytes')
+            self.cmd('extract', self.repository_location + '::test')
+        self.assert_dirs_equal('input', 'output/input')
+
     def test_extract_list_output(self):
         self.cmd('init', self.repository_location)
         self.create_regular_file('file', size=1024 * 80)
diff --git a/src/borg/testsuite/helpers.py b/src/borg/testsuite/helpers.py
index 55569e96e1..c9656c71ad 100644
--- a/src/borg/testsuite/helpers.py
+++ b/src/borg/testsuite/helpers.py
@@ -10,7 +10,8 @@
 import msgpack.fallback
 
 from .. import platform
-from ..helpers import Location
+from ..cache import ChunkListEntry
+from ..helpers import Location, slice_chunks
 from ..helpers import Buffer
 from ..helpers import partial_format, format_file_size, parse_file_size, format_timedelta, format_line, PlaceholderError, replace_placeholders
 from ..helpers import make_path_safe, clean_lines
@@ -1081,3 +1082,29 @@ def test_swidth_slice_mixed_characters():
     string = '나윤a선나윤선나윤선나윤선나윤선'
     assert swidth_slice(string, 5) == '나윤a'
     assert swidth_slice(string, 6) == '나윤a'
+
+
+class TestSliceChunks:
+    @staticmethod
+    def chunk(size):
+        return ChunkListEntry(None, size, 0)
+
+    @pytest.mark.parametrize('chunks, offset, expected_chunks, expected_prefix_length', (
+        # Edge case: offset exactly on chunk boundary
+        ([(1, 1000), (2, 500)], 1000, [(2, 500)], 1000),
+        ([(1, 1000), (2, 500)], 999, [(1, 1000), (2, 500)], 0),
+        ([(1, 1000), (2, 500)], 1001, [(2, 500)], 1000),
+
+        # Edge case: offset completely consumes
+        ([(1, 1000), (2, 500)], 1500, [], 1500),
+        ([(1, 1000), (2, 500)], 1499, [(2, 500)], 1000),
+
+        # Edge case: offset > length of chunks
+        ([(1, 1000), (2, 500)], 1501, [], 1500),
+    ))
+    def test_basic(self, chunks, offset, expected_chunks, expected_prefix_length):
+        chunks = [ChunkListEntry(id, size, 0) for id, size in chunks]
+        expected_chunks = [ChunkListEntry(id, size, 0) for id, size in expected_chunks]
+        chunks, remaing_offset = slice_chunks(chunks, offset)
+        assert chunks == expected_chunks
+        assert remaing_offset == expected_prefix_length