diff --git a/evals/metrics/ragaaf/prompt_templates/context_recall.py b/evals/metrics/ragaaf/prompt_templates/context_recall.py
new file mode 100644
index 00000000..fbbb6dbe
--- /dev/null
+++ b/evals/metrics/ragaaf/prompt_templates/context_recall.py
@@ -0,0 +1,13 @@
+# Copyright (C) 2024 Intel Corporation
+# SPDX-License-Identifier: Apache-2.0
+
+
+class ContextRecall:
+    name = "context_recall"
+    required_columns = ["context", "ground_truth"]
+    template = """- Context Recall: context recall measures how well ground truth is supported by the context.
+  - Score 1: None of facts/steps/aspects in the ground truth are supported by the context. 
+  - Score 2: Only a small part of facts/steps/aspects in the ground truth are supported by the context. 
+  - Score 3: About half of facts/steps/aspects in the ground truth are supported by the context. 
+  - Score 4: All main/important facts/steps/aspects in the ground truth are supported by the context but some of the details are not supported by the context. 
+  - Score 5: All of facts/steps/aspects in the ground truth are supported by the context. """