diff --git a/optillm/__init__.py b/optillm/__init__.py
index 32c5c31..e6c30c8 100644
--- a/optillm/__init__.py
+++ b/optillm/__init__.py
@@ -1,5 +1,5 @@
 # Version information
-__version__ = "0.3.18"
+__version__ = "0.3.19"
 
 import os as _os
 
diff --git a/pyproject.toml b/pyproject.toml
index 99e9edb..6ac8d1b 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "optillm"
-version = "0.3.18"
+version = "0.3.19"
 description = "An optimizing inference proxy for LLMs."
 readme = "README.md"
 license = "Apache-2.0"
diff --git a/requirements.txt b/requirements.txt
index e262ae2..04b2181 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -5,7 +5,11 @@ z3-solver<=4.15.4.0
 aiohttp
 flask
 torch
-transformers>=5.0.0
+# Cap below 5.13: transformers 5.13.0 tightened AutoTokenizer.register() to require
+# a class, which breaks mlx-lm (it registers a tokenizer by string name). Lift the
+# cap once mlx-lm ships a transformers-5.13-compatible release. See mlx_lm
+# tokenizer_utils.register("NewlineTokenizer", ...).
+transformers>=5.0.0,<5.13.0
 azure.identity
 tiktoken
 scikit-learn