diff --git a/pyproject.toml b/pyproject.toml
index ebb62207b..c22f04f28 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -414,21 +414,23 @@ basepython = "pypy3.11"
 
 [tool.tox.env.ruff]
 basepython = "python"
-deps = "ruff"
 commands = [["ruff", "check", "pythainlp"], ["ruff", "format", "--check", "pythainlp"]]
+deps = "ruff"
 skip_install = true
 
 # Mypy configuration
 [tool.mypy]
 check_untyped_defs = true
-disallow_untyped_defs = false
-disallow_incomplete_defs = false
-disallow_untyped_decorators = false
+disallow_incomplete_defs = true
+disallow_subclassing_any = true
+disallow_untyped_decorators = true
+disallow_untyped_defs = true
 pretty = true
 python_version = "3.9"
 show_column_numbers = true
 show_error_code_links = true
 show_error_context = true
+strict = true
 strict_bytes = true
 strict_equality = true
 strict_equality_for_none = true
@@ -436,18 +438,9 @@ strict_optional = true
 warn_no_return = true
 warn_redundant_casts = true
 warn_return_any = true
+warn_unreachable = true
 warn_unused_configs = true
 warn_unused_ignores = true
-warn_unreachable = true
-
-# Per-module options for stricter checking on foundation modules
-[[tool.mypy.overrides]]
-module = [
-    "pythainlp.tools.*",
-    "pythainlp.cli.*",
-]
-disallow_untyped_defs = true
-disallow_incomplete_defs = true
 
 # Ignore missing imports for optional dependencies
 [[tool.mypy.overrides]]
diff --git a/pythainlp/spell/wanchanberta_thai_grammarly.py b/pythainlp/spell/wanchanberta_thai_grammarly.py
index 2846df4e2..5ac7003f1 100644
--- a/pythainlp/spell/wanchanberta_thai_grammarly.py
+++ b/pythainlp/spell/wanchanberta_thai_grammarly.py
@@ -31,7 +31,7 @@
 )
 
 
-class BertModel(torch.nn.Module):
+class BertModel(torch.nn.Module):  # type: ignore[misc]
     def __init__(self) -> None:
         super().__init__()
         self.bert: BertForTokenClassification = (
diff --git a/pythainlp/translate/tokenization_small100.py b/pythainlp/translate/tokenization_small100.py
index 6625cf9a5..e40cbd1a4 100644
--- a/pythainlp/translate/tokenization_small100.py
+++ b/pythainlp/translate/tokenization_small100.py
@@ -65,7 +65,7 @@
 # fmt: on
 
 
-class SMALL100Tokenizer(PreTrainedTokenizer):
+class SMALL100Tokenizer(PreTrainedTokenizer):  # type: ignore[misc]
     """Construct an SMALL100 tokenizer. Based on [SentencePiece](https://github.com/google/sentencepiece).
     This tokenizer inherits from [`PreTrainedTokenizer`] which contains most of the main methods. Users should refer to
     this superclass for more information regarding those methods.
diff --git a/pythainlp/transliterate/thai2rom.py b/pythainlp/transliterate/thai2rom.py
index 39899bb5f..14ce6ebc5 100644
--- a/pythainlp/transliterate/thai2rom.py
+++ b/pythainlp/transliterate/thai2rom.py
@@ -119,7 +119,7 @@ def romanize(self, text: str) -> str:
         return "".join(target)
 
 
-class Encoder(nn.Module):
+class Encoder(nn.Module):  # type: ignore[misc]
     hidden_size: int
     character_embedding: nn.Embedding
     rnn: nn.LSTM
@@ -196,7 +196,7 @@ def init_hidden(
         return (h_0, c_0)
 
 
-class Attn(nn.Module):
+class Attn(nn.Module):  # type: ignore[misc]
     method: str
     hidden_size: int
     attn: nn.Linear
@@ -256,7 +256,7 @@ def forward(
         return F.softmax(attn_energies, 1)
 
 
-class AttentionDecoder(nn.Module):
+class AttentionDecoder(nn.Module):  # type: ignore[misc]
     vocabulary_size: int
     hidden_size: int
     character_embedding: nn.Embedding
@@ -324,7 +324,7 @@ def forward(
         return x, hidden[0], attn_weights
 
 
-class Seq2Seq(nn.Module):
+class Seq2Seq(nn.Module):  # type: ignore[misc]
     encoder: Encoder
     decoder: AttentionDecoder
     pad_idx: int
diff --git a/pythainlp/transliterate/thaig2p.py b/pythainlp/transliterate/thaig2p.py
index 7813e801e..f1438198d 100644
--- a/pythainlp/transliterate/thaig2p.py
+++ b/pythainlp/transliterate/thaig2p.py
@@ -131,7 +131,7 @@ def g2p(self, text: str) -> str:
         return "".join(target)
 
 
-class Encoder(nn.Module):
+class Encoder(nn.Module):  # type: ignore[misc]
     hidden_size: int
     character_embedding: nn.Embedding
     rnn: nn.LSTM
@@ -216,7 +216,7 @@ def init_hidden(
         return (h_0, c_0)
 
 
-class Attn(nn.Module):
+class Attn(nn.Module):  # type: ignore[misc]
     method: str
     hidden_size: int
     attn: nn.Linear
@@ -278,7 +278,7 @@ def forward(
         return F.softmax(attn_energies, 1)
 
 
-class AttentionDecoder(nn.Module):
+class AttentionDecoder(nn.Module):  # type: ignore[misc]
     vocabulary_size: int
     hidden_size: int
     character_embedding: nn.Embedding
@@ -346,7 +346,7 @@ def forward(
         return x, hidden[0], attn_weights
 
 
-class Seq2Seq(nn.Module):
+class Seq2Seq(nn.Module):  # type: ignore[misc]
     encoder: Encoder
     decoder: AttentionDecoder
     pad_idx: int