diff --git a/src/prompt_toolkit/input/vt100_parser.py b/src/prompt_toolkit/input/vt100_parser.py
index 34ea11057..6a8cd8bd6 100644
--- a/src/prompt_toolkit/input/vt100_parser.py
+++ b/src/prompt_toolkit/input/vt100_parser.py
@@ -188,6 +188,12 @@ def _call_handler(
                 self._in_bracketed_paste = True
                 self._paste_buffer = ""
             else:
+                # For character-valued ANSI_SEQUENCES entries (plain characters
+                # that are not Keys members), use the character as insert_text
+                # instead of the raw escape sequence so that self-insert
+                # bindings work correctly.
+                if not isinstance(key, Keys):
+                    insert_text = key
                 self.feed_key_callback(KeyPress(key, insert_text))
 
     def feed(self, data: str) -> None:
diff --git a/tests/test_inputstream.py b/tests/test_inputstream.py
index ab1b03689..63b5deb0f 100644
--- a/tests/test_inputstream.py
+++ b/tests/test_inputstream.py
@@ -139,3 +139,26 @@ def test_cpr_response_2(processor, stream):
     assert len(processor.keys) == 2
     assert processor.keys[0].key == Keys.CPRResponse
     assert processor.keys[1].key == Keys.ControlJ
+
+
+def test_character_valued_ansi_sequence(processor, stream):
+    """Character-valued ANSI_SEQUENCES entries should use the character as data.
+
+    Regression test for: https://github.com/prompt-toolkit/python-prompt-toolkit/issues/2086
+    When a character-valued entry is added to ANSI_SEQUENCES (e.g., for xterm
+    modifyOtherKeys support), the KeyPress data should be the character, not the
+    raw escape sequence.
+    """
+    from prompt_toolkit.input.ansi_escape_sequences import ANSI_SEQUENCES
+
+    SEQ = "\x1b[27;2;78~"  # Shift+N via xterm modifyOtherKeys=2
+    ANSI_SEQUENCES[SEQ] = "N"
+
+    try:
+        stream.feed(SEQ)
+        assert len(processor.keys) == 1
+        assert processor.keys[0].key == "N"
+        # data should be the character, not the raw escape sequence
+        assert processor.keys[0].data == "N"
+    finally:
+        del ANSI_SEQUENCES[SEQ]