diff --git a/README.md b/README.md
index 95009f5..3e4e3cc 100644
--- a/README.md
+++ b/README.md
@@ -137,6 +137,14 @@ def foo() -> tuple[dict[str, int], str | None]:
 
 Unused import names will be removed, and if `from __future__ import annotations` is not found in the script, it will be automatically added if the new syntax is being used.
 
+## Use as a command line tool
+
+```bash
+python3 -m pip install -U fix-future-annotations
+
+fix-future-annotations my_script.py
+```
+
 ## Use as pre-commit hook
 
 Add the following to your `.pre-commit-config.yaml`:
@@ -149,12 +157,21 @@ repos:
       - id: fix-future-annotations
 ```
 
-## Use as command line tool
+## Configurations
 
-```bash
-python3 -m pip install -U fix-future-annotations
+`fix-future-annotations` can be configured via `pyproject.toml`. Here is an example:
 
-fix-future-annotations my_script.py
+```toml
+[tool.fix_future_annotations]
+exclude_files = [  # regex patterns to exclude files
+    'tests/.*',
+    'docs/.*',
+]
+
+exclude_lines = [  # regex patterns to exclude lines
+    '# ffa: ignore',   # if a line ends with this comment, the whole *block* will be excluded
+    'class .+\(BaseModel\):'  # classes that inherit from `BaseModel` will be excluded
+]
 ```
 
 ## License
diff --git a/fix_future_annotations/__init__.py b/fix_future_annotations/__init__.py
index e69de29..02866c3 100644
--- a/fix_future_annotations/__init__.py
+++ b/fix_future_annotations/__init__.py
@@ -0,0 +1,4 @@
+from fix_future_annotations._main import fix_file
+
+
+__all__ = ["fix_file"]
diff --git a/fix_future_annotations/_config.py b/fix_future_annotations/_config.py
new file mode 100644
index 0000000..2deaa82
--- /dev/null
+++ b/fix_future_annotations/_config.py
@@ -0,0 +1,38 @@
+from __future__ import annotations
+
+from dataclasses import dataclass, field
+from pathlib import Path
+import re
+import sys
+
+if sys.version_info >= (3, 11):
+    import tomllib
+else:
+    import tomli as tomllib
+
+
+@dataclass
+class Config:
+    """Configuration for fix_future_annotations."""
+
+    # The line patterns(regex) to exclude from the fix.
+    exclude_lines: list[str] = field(default_factory=list)
+    # The file patterns(regex) to exclude from the fix.
+    exclude_files: list[str] = field(default_factory=list)
+
+    @classmethod
+    def from_file(cls, path: str | Path = "pyproject.toml") -> Config:
+        """Load the configuration from a file."""
+        try:
+            with open(path, "rb") as f:
+                data = tomllib.load(f)
+        except OSError:
+            return cls()
+        else:
+            return cls(**data.get("tool", {}).get("fix_future_annotations", {}))
+
+    def is_file_excluded(self, file_path: str) -> bool:
+        return any(re.search(pattern, file_path) for pattern in self.exclude_files)
+
+    def is_line_excluded(self, line: str) -> bool:
+        return any(re.search(pattern, line) for pattern in self.exclude_lines)
diff --git a/fix_future_annotations/_main.py b/fix_future_annotations/_main.py
index 71fa903..7706b3b 100644
--- a/fix_future_annotations/_main.py
+++ b/fix_future_annotations/_main.py
@@ -10,6 +10,7 @@
 
 from tokenize_rt import reversed_enumerate, src_to_tokens, tokens_to_src
 
+from fix_future_annotations._config import Config
 from fix_future_annotations._visitor import AnnotationVisitor
 
 
@@ -17,17 +18,19 @@ def _escaped(line: str) -> bool:
     return (len(line) - len(line.rstrip("\\"))) % 2 == 1
 
 
-def _iter_files(*paths: str) -> Iterator[str]:
+def _iter_files(*paths: str, config: Config) -> Iterator[str]:
     def files_under_dir(path: str) -> Iterator[str]:
         for root, _, files in os.walk(path):
             for filename in files:
                 if filename.endswith(".py"):
-                    yield os.path.join(root, filename)
+                    fn = os.path.join(root, filename).replace("\\", "/")
+                    if not config.is_file_excluded(fn):
+                        yield fn
 
     for path in paths:
         if os.path.isdir(path):
             yield from files_under_dir(path)
-        elif path.endswith(".py"):
+        elif path.endswith(".py") and not config.is_file_excluded(path):
             yield path
 
 
@@ -82,14 +85,20 @@ def _add_future_annotations(content: str) -> str:
 
 
 def fix_file(
-    file_path: str | Path, write: bool = False, show_diff: bool = False
+    file_path: str | Path,
+    *,
+    write: bool = False,
+    show_diff: bool = False,
+    config: Config | None = None,
 ) -> bool:
     """Fix the file at file_path to use PEP 585, 604 and 563 syntax."""
+    if config is None:
+        config = Config.from_file()
     file_path = Path(file_path)
     file_content = file_path.read_text("utf-8")
     tokens = src_to_tokens(file_content)
     tree = ast.parse(file_content)
-    visitor = AnnotationVisitor()
+    visitor = AnnotationVisitor(file_content.splitlines(), config=config)
     token_funcs = visitor.get_token_functions(tree)
     for i, token in reversed_enumerate(tokens):
         if not token.src:
@@ -137,9 +146,12 @@ def main(argv: list[str] | None = None) -> None:
     args = parser.parse_args(argv)
     diff_count = 0
     checked = 0
-    for filename in _iter_files(*args.path):
+    config = Config.from_file()
+    for filename in _iter_files(*args.path, config=config):
         checked += 1
-        result = fix_file(filename, args.write, show_diff=args.verbose)
+        result = fix_file(
+            filename, write=args.write, show_diff=args.verbose, config=config
+        )
         diff_count += int(result)
     if diff_count:
         if args.write:
diff --git a/fix_future_annotations/_visitor.py b/fix_future_annotations/_visitor.py
index 21091f6..f5c914d 100644
--- a/fix_future_annotations/_visitor.py
+++ b/fix_future_annotations/_visitor.py
@@ -8,6 +8,7 @@
 
 from tokenize_rt import NON_CODING_TOKENS, Offset, Token
 
+from fix_future_annotations._config import Config
 from fix_future_annotations._utils import (
     ast_to_offset,
     find_closing_bracket,
@@ -134,11 +135,17 @@ def _fix_union(i: int, tokens: list[Token], *, arg_count: int) -> None:
 class State(NamedTuple):
     in_annotation: bool
     in_literal: bool
+    omit: bool
+
+    def update_annotation(self) -> bool:
+        return self.in_annotation and not self.omit
 
 
 class AnnotationVisitor(ast.NodeVisitor):
-    def __init__(self) -> None:
+    def __init__(self, lines: list[str], *, config: Config) -> None:
         super().__init__()
+        self.lines = lines
+        self.config = config
         self.token_funcs: dict[Offset, list[TokenFunc]] = {}
 
         self._typing_import_name: str | None = None
@@ -162,8 +169,12 @@ def add_conditional_token_func(
             (condition, partial(self.add_token_func, offset, func))
         )
 
+    def _is_excluded(self, node: ast.AST) -> bool:
+        line = self.lines[node.lineno - 1]
+        return self.config.is_line_excluded(line)
+
     def get_token_functions(self, tree: ast.Module) -> dict[Offset, list[TokenFunc]]:
-        with self.under_state(State(False, False)):
+        with self.under_state(State(False, False, False)):
             self.visit(tree)
         for condition, callback in self._conditional_callbacks:
             if condition():
@@ -188,6 +199,14 @@ def under_state(self, state: State) -> None:
         finally:
             self._state_stack.pop()
 
+    def visit(self, node: ast.AST) -> Any:
+        if isinstance(node, ast.stmt) and self._is_excluded(node):
+            ctx = self.under_state(self.state._replace(omit=True))
+        else:
+            ctx = contextlib.nullcontext()
+        with ctx:
+            return super().visit(node)
+
     def generic_visit(self, node: ast.AST) -> Any:
         for field in reversed(node._fields):
             value = getattr(node, field)
@@ -244,7 +263,7 @@ def visit_ImportFrom(self, node: ast.ImportFrom) -> Any:
     def visit_Attribute(self, node: ast.Attribute) -> Any:
         """Transform typing.List -> list"""
         if (
-            self.state.in_annotation
+            self.state.update_annotation()
             and isinstance(node.value, ast.Name)
             and node.value.id == self._typing_import_name
             and node.attr in BASIC_COLLECTION_TYPES
@@ -258,7 +277,7 @@ def visit_Attribute(self, node: ast.Attribute) -> Any:
     def visit_Name(self, node: ast.Name) -> Any:
         if node.id in self._typing_imports_to_remove:
             name = self._typing_imports_to_remove[node.id]
-            if not self.state.in_annotation:
+            if not self.state.update_annotation():
                 # It is referred to outside of an annotation, so we need to exclude it
                 self._conditional_callbacks.insert(
                     0,
@@ -283,7 +302,7 @@ def visit_BinOp(self, node: ast.BinOp) -> Any:
         return self.generic_visit(node)
 
     def visit_Subscript(self, node: ast.Subscript) -> Any:
-        if not self.state.in_annotation:
+        if not self.state.update_annotation():
             return self.generic_visit(node)
         if isinstance(node.value, ast.Attribute):
             if (
@@ -327,7 +346,7 @@ def visit_Subscript(self, node: ast.Subscript) -> Any:
 
     def visit_Constant(self, node: ast.Constant) -> Any:
         if (
-            self.state.in_annotation
+            self.state.update_annotation()
             and not self.state.in_literal
             and isinstance(node.value, str)
         ):
diff --git a/pdm.lock b/pdm.lock
index f8b37e8..60a67c1 100644
--- a/pdm.lock
+++ b/pdm.lock
@@ -70,8 +70,8 @@ requires_python = ">=3.7"
 summary = "A lil' TOML parser"
 
 [metadata]
-lock_version = "4.0"
-content_hash = "sha256:13ed22f7e12e7c7c3ee7cdeb18d0a3764932cac684d130d177be14ad0546d577"
+lock_version = "4.1"
+content_hash = "sha256:b50b8c49b0bcf0f0b3c1d684310b54ec461bdd90ce8804ff7e57f17092d2e57b"
 
 [metadata.files]
 "attrs 22.1.0" = [
diff --git a/pyproject.toml b/pyproject.toml
index e93ed49..f365fac 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -6,6 +6,7 @@ authors = [
 ]
 dependencies = [
     "tokenize-rt>=5.0.0",
+    "tomli; python_version < '3.11'",
 ]
 requires-python = ">=3.8"
 readme = "README.md"
@@ -49,3 +50,9 @@ exclude = '''
   | tests/samples
 )/
 '''
+
+
+[tool.fix_future_annotations]
+exclude_lines = [
+    "# ffa: ignore"
+]
diff --git a/tests/samples/exclude_lines.py b/tests/samples/exclude_lines.py
new file mode 100644
index 0000000..c42eabb
--- /dev/null
+++ b/tests/samples/exclude_lines.py
@@ -0,0 +1,19 @@
+from typing import List, Optional, Tuple, Union
+
+
+class NoFix:
+    def __init__(self, names: List[str]) -> None:
+        self.names = names
+
+    def lengh(self) -> Optional[int]:
+        if self.names:
+            return len(self.names)
+        return None
+
+
+def foo() -> Union[str, int]:  # ffa: ignore
+    return 42
+
+
+def bar() -> Tuple[str, int]:
+    return "bar", 42
diff --git a/tests/samples/exclude_lines_fix.py b/tests/samples/exclude_lines_fix.py
new file mode 100644
index 0000000..b6be68b
--- /dev/null
+++ b/tests/samples/exclude_lines_fix.py
@@ -0,0 +1,21 @@
+from __future__ import annotations
+
+from typing import List, Optional, Union
+
+
+class NoFix:
+    def __init__(self, names: List[str]) -> None:
+        self.names = names
+
+    def lengh(self) -> Optional[int]:
+        if self.names:
+            return len(self.names)
+        return None
+
+
+def foo() -> Union[str, int]:  # ffa: ignore
+    return 42
+
+
+def bar() -> tuple[str, int]:
+    return "bar", 42
diff --git a/tests/test_fix_future_annotations.py b/tests/test_fix_future_annotations.py
index 5190bde..786565f 100644
--- a/tests/test_fix_future_annotations.py
+++ b/tests/test_fix_future_annotations.py
@@ -3,6 +3,7 @@
 import pytest
 
 from fix_future_annotations._main import fix_file
+from fix_future_annotations._config import Config
 
 SAMPLES = Path(__file__).with_name("samples")
 
@@ -19,9 +20,10 @@ def _load_samples() -> list:
 @pytest.mark.parametrize("origin, fixed", _load_samples())
 def test_fix_samples(origin: Path, fixed: Path, tmp_path: Path) -> None:
     copied = shutil.copy2(origin, tmp_path)
-    result = fix_file(copied, True)
+    config = Config(exclude_lines=["# ffa: ignore", "class NoFix:"])
+    result = fix_file(copied, write=True, config=config)
 
     assert fixed.read_text() == Path(copied).read_text()
 
-    result = fix_file(copied, False)
+    result = fix_file(copied, write=False, config=config)
     assert not result