diff --git a/CHANGELOG.md b/CHANGELOG.md
index c7fbaa7..26796d7 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -7,8 +7,14 @@ and the project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.
 
 ## [Unreleased]
 
+### Changed
+
 - Drop the explicit TensorBoard dependency ([#12](https://github.com/microsoft/retrochimera/pull/12)) ([@kmaziarz])
 
+### Added
+
+- Expose setting `num_processes` for template-based models ([#13](https://github.com/microsoft/retrochimera/pull/13)) ([@kmaziarz])
+
 ## [1.1.0] - 2026-03-12
 
 ### Changed
diff --git a/retrochimera/chem/rules.py b/retrochimera/chem/rules.py
index ccfe82c..2b821f3 100644
--- a/retrochimera/chem/rules.py
+++ b/retrochimera/chem/rules.py
@@ -182,6 +182,7 @@ def __init__(
         max_cumulative_prob: float = 1.0,
         apply_rules_timeout: Optional[float] = None,
         include_all_metadata: bool = False,
+        num_processes: Optional[int] = None,
         **kwargs,
     ) -> None:
         """Initialize a rule-based model.
@@ -199,6 +200,8 @@ def __init__(
                 not set, we fallback to the default setting in the server.
             include_all_metadata: If set, model outputs will include detailed metadata (may require
                 substantial disk space for saving the results).
+            num_processes: Number of parallel processes for the rule application server. If not set,
+                defaults to the number of available CPUs.
         """
         super().__init__(**kwargs)  # In case this is one of several base classes in the MRO.
 
@@ -208,6 +211,7 @@ def __init__(
         self._max_cumulative_prob = max_cumulative_prob
         self._apply_rules_timeout = apply_rules_timeout
         self._include_all_metadata = include_all_metadata
+        self._num_processes = num_processes
 
     def start_server(self, rulebase_dir: Union[str, Path]) -> None:
         """Instantiate the rule application server lazily."""
@@ -215,9 +219,15 @@ def start_server(self, rulebase_dir: Union[str, Path]) -> None:
         # Local to avoid circular import.
         from retrochimera.chem.rule_application_server import RuleApplicationServer
 
-        self._server = RuleApplicationServer(
-            rulebase_dir=rulebase_dir, rule_application_kwargs=self._rule_application_server_kwargs
-        )
+        server_kwargs: dict[str, Any] = {
+            "rulebase_dir": rulebase_dir,
+            "rule_application_kwargs": self._rule_application_server_kwargs,
+        }
+
+        if self._num_processes is not None:
+            server_kwargs["num_processes"] = self._num_processes
+
+        self._server = RuleApplicationServer(**server_kwargs)
 
     def _get_server(self):
         if self._server is None: