diff --git a/python/embeddings.py b/python/embeddings.py
new file mode 100644
index 000000000..ba92f328d
--- /dev/null
+++ b/python/embeddings.py
@@ -0,0 +1,62 @@
+import mgp
+import subprocess
+import sys
+
+
+EXCLUDE_PROPERTIES = {"embedding"}
+
+logger: mgp.Logger = mgp.Logger()
+
+@mgp.write_proc
+def compute_embeddings(ctx: mgp.ProcCtx, node: mgp.Vertex) -> mgp.Record(embedding_string=str, success=bool):
+
+    try:
+        from sentence_transformers import SentenceTransformer
+    except ImportError:
+        # Make sure pip is there
+        try:
+            subprocess.check_call([sys.executable, "ensurepip"])
+        except subprocess.CalledProcessError:
+            logger.error("Failed to ensure pip is available")
+            return mgp.Record(embedding_string="", success=False)
+        
+        # Install the sentence-transformers package
+        try:
+            subprocess.check_call([
+                sys.executable, "pip", "install", "sentence-transformers"
+            ])
+            from sentence_transformers import SentenceTransformer
+        except subprocess.CalledProcessError:
+            logger.error("Failed to install the sentence-transformers package")
+            return mgp.Record(embedding_string="", success=False)
+
+    try: 
+        model = SentenceTransformer("all-MiniLM-L6-v2")
+    except Exception as e:
+        logger.error(f"Failed to load model: {e}")
+        return mgp.Record(embedding_string="", success=False)
+
+    try:
+
+        for vertex in ctx.graph.vertices:
+
+            # Test id: 555 name: Pero last_name: Peric nums: (1, 2, 3) birthday: 1947-07-30 maps: {'day': 30, 'month': 7, 'year': 1947} lap: 0:02:02.000033
+
+            #TODO: parametrize the exluded properties
+            node_data = " ".join(label.name for label in vertex.labels) + " " + " ".join(
+                f"{key}: {value}"
+                for key, value in vertex.properties.items()
+                if key not in EXCLUDE_PROPERTIES
+            )
+            # Compute the embedding
+            node_embedding = model.encode(node_data)
+
+            #TODO: parametrize the property name
+            vertex.properties["embedding"] = node_embedding.tolist()
+
+        return mgp.Record(embedding_string=node_data, success=True)
+
+    except Exception as e:
+        # Handle exceptions by returning failure status
+        logger.error(f"Failed to compute embedding for node: {e}")
+        return mgp.Record(embedding_string="", success=False)