diff --git a/fmtr/tools/semantic_tools.py b/fmtr/tools/semantic_tools.py
new file mode 100644
index 0000000..be2bf2f
--- /dev/null
+++ b/fmtr/tools/semantic_tools.py
@@ -0,0 +1,52 @@
+import pandas as pd
+from sentence_transformers import SentenceTransformer
+
+from fmtr.tools.logging_tools import logger
+
+
+class SemanticManager:
+    """
+
+    Base semantic similarity manager
+
+    """
+
+    REPO_ID = 'distiluse-base-multilingual-cased-v1'
+
+    def __init__(self, data: pd.Series):
+        logger.info(f"Loading model from {self.REPO_ID}")
+        self.model = SentenceTransformer(self.REPO_ID)
+        self.data = data
+        logger.info(f"Vectorising {len(data)} texts using {self.model.device}...")
+        self.embs = self.vectorise()
+
+    def vectorise(self):
+        """
+
+        Vectorise the corpus
+
+        """
+        embs = self.model.encode(self.data.tolist())
+        return embs
+
+    def get_sims(self, string: str):
+        """
+
+        Get similarities between query string and corpus
+
+        """
+        logger.info(f'Getting similarities for search term: "{string}"...')
+        embs_query = self.model.encode([string])
+        sims = self.model.similarity(self.embs, embs_query).squeeze().numpy()
+        return sims
+
+    def get_matches(self, string: str, top_n: int = 20):
+        """
+
+        Get the Top N matches between query string and corpus
+
+        """
+        sims = self.get_sims(string)
+        args = sims.argsort()[::-1]
+        matches = self.data.iloc[args][:top_n]
+        return matches