diff --git a/CHANGELOG.md b/CHANGELOG.md
index af4997a..1ef068c 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -4,6 +4,7 @@
 1. Add a new get_relationship_properties_keys tool.
 2. Add targetNode filtering for longest_path.
 3. Add support for similarity algorithms.
+4. Add relationship directionality used by graph projection as an optional parameter to all appropriate algorithms. This include all algorithms that support both directed and undirected graphs and behave differently.
 
 ### Bug Fixes
 1. Return node names in several path algorithms that only returned node ids.
diff --git a/mcp_server/src/mcp_server_neo4j_gds/centrality_algorithm_handlers.py b/mcp_server/src/mcp_server_neo4j_gds/centrality_algorithm_handlers.py
index 97f7892..716b7df 100644
--- a/mcp_server/src/mcp_server_neo4j_gds/centrality_algorithm_handlers.py
+++ b/mcp_server/src/mcp_server_neo4j_gds/centrality_algorithm_handlers.py
@@ -13,8 +13,8 @@
 
 
 class ArticleRankHandler(AlgorithmHandler):
-    def article_rank(self, **kwargs):
-        with projected_graph(self.gds) as G:
+    def article_rank(self, undirected: bool = False, **kwargs):
+        with projected_graph(self.gds, undirected=undirected) as G:
             # If any optional parameter is not None, use that parameter
             args = locals()
             params = {
@@ -46,6 +46,7 @@ def article_rank(self, **kwargs):
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.article_rank(
+            undirected=arguments.get("undirected", False),
             nodes=arguments.get("nodes"),
             nodeIdentifierProperty=arguments.get("nodeIdentifierProperty"),
             sourceNodes=arguments.get("sourceNodes"),
@@ -75,8 +76,8 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
 
 
 class BetweennessCentralityHandler(AlgorithmHandler):
-    def betweenness_centrality(self, **kwargs):
-        with projected_graph(self.gds) as G:
+    def betweenness_centrality(self, undirected: bool = False, **kwargs):
+        with projected_graph(self.gds, undirected=undirected) as G:
             params = {
                 k: v
                 for k, v in kwargs.items()
@@ -99,6 +100,7 @@ def betweenness_centrality(self, **kwargs):
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.betweenness_centrality(
+            undirected=arguments.get("undirected", False),
             nodes=arguments.get("nodes"),
             nodeIdentifierProperty=arguments.get("nodeIdentifierProperty"),
             samplingSize=arguments.get("samplingSize"),
@@ -124,13 +126,13 @@ def bridges(self, **kwargs):
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.bridges(
-            nodeIdentifierProperty=arguments.get("nodeIdentifierProperty")
+            nodeIdentifierProperty=arguments.get("nodeIdentifierProperty"),
         )
 
 
 class CELFHandler(AlgorithmHandler):
-    def celf(self, **kwargs):
-        with projected_graph(self.gds) as G:
+    def celf(self, undirected: bool = False, **kwargs):
+        with projected_graph(self.gds, undirected=undirected) as G:
             params = {
                 k: v
                 for k, v in kwargs.items()
@@ -147,6 +149,7 @@ def celf(self, **kwargs):
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.celf(
+            undirected=arguments.get("undirected", False),
             seedSetSize=arguments.get("seedSetSize"),
             monteCarloSimulations=arguments.get("monteCarloSimulations"),
             propagationProbability=arguments.get("propagationProbability"),
@@ -155,8 +158,8 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
 
 
 class ClosenessCentralityHandler(AlgorithmHandler):
-    def closeness_centrality(self, **kwargs):
-        with projected_graph(self.gds) as G:
+    def closeness_centrality(self, undirected: bool = False, **kwargs):
+        with projected_graph(self.gds, undirected=undirected) as G:
             params = {
                 k: v
                 for k, v in kwargs.items()
@@ -174,11 +177,11 @@ def closeness_centrality(self, **kwargs):
         centrality = filter_identifiers(
             self.gds, node_identifier_property, node_names, centrality
         )
-
         return centrality
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.closeness_centrality(
+            undirected=arguments.get("undirected", False),
             nodes=arguments.get("nodes"),
             nodeIdentifierProperty=arguments.get("nodeIdentifierProperty"),
             useWassermanFaust=arguments.get("useWassermanFaust"),
@@ -186,8 +189,8 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
 
 
 class DegreeCentralityHandler(AlgorithmHandler):
-    def degree_centrality(self, **kwargs):
-        with projected_graph(self.gds) as G:
+    def degree_centrality(self, undirected: bool = False, **kwargs):
+        with projected_graph(self.gds, undirected=undirected) as G:
             params = {
                 k: v
                 for k, v in kwargs.items()
@@ -210,6 +213,7 @@ def degree_centrality(self, **kwargs):
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.degree_centrality(
+            undirected=arguments.get("undirected", False),
             nodes=arguments.get("nodes"),
             nodeIdentifierProperty=arguments.get("nodeIdentifierProperty"),
             orientation=arguments.get("orientation"),
@@ -217,8 +221,8 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
 
 
 class EigenvectorCentralityHandler(AlgorithmHandler):
-    def eigenvector_centrality(self, **kwargs):
-        with projected_graph(self.gds) as G:
+    def eigenvector_centrality(self, undirected: bool = False, **kwargs):
+        with projected_graph(self.gds, undirected=undirected) as G:
             params = {
                 k: v
                 for k, v in kwargs.items()
@@ -250,6 +254,7 @@ def eigenvector_centrality(self, **kwargs):
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.eigenvector_centrality(
+            undirected=arguments.get("undirected", False),
             nodes=arguments.get("nodes"),
             nodeIdentifierProperty=arguments.get("nodeIdentifierProperty"),
             maxIterations=arguments.get("maxIterations"),
@@ -261,8 +266,8 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
 
 
 class PageRankHandler(AlgorithmHandler):
-    def pagerank(self, **kwargs):
-        with projected_graph(self.gds) as G:
+    def pagerank(self, undirected: bool = False, **kwargs):
+        with projected_graph(self.gds, undirected=undirected) as G:
             params = {
                 k: v
                 for k, v in kwargs.items()
@@ -293,6 +298,7 @@ def pagerank(self, **kwargs):
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.pagerank(
+            undirected=arguments.get("undirected", False),
             nodes=arguments.get("nodes"),
             nodeIdentifierProperty=arguments.get("nodeIdentifierProperty"),
             sourceNodes=arguments.get("sourceNodes"),
@@ -303,8 +309,8 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
 
 
 class HarmonicCentralityHandler(AlgorithmHandler):
-    def harmonic_centrality(self, **kwargs):
-        with projected_graph(self.gds) as G:
+    def harmonic_centrality(self, undirected: bool = False, **kwargs):
+        with projected_graph(self.gds, undirected=undirected) as G:
             centrality = self.gds.closeness.harmonic.stream(G)
 
         # Add node names to the results if nodeIdentifierProperty is provided
@@ -320,14 +326,15 @@ def harmonic_centrality(self, **kwargs):
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.harmonic_centrality(
+            undirected=arguments.get("undirected", False),
             nodes=arguments.get("nodes"),
             nodeIdentifierProperty=arguments.get("nodeIdentifierProperty"),
         )
 
 
 class HITSHandler(AlgorithmHandler):
-    def hits(self, **kwargs):
-        with projected_graph(self.gds) as G:
+    def hits(self, undirected: bool = False, **kwargs):
+        with projected_graph(self.gds, undirected=undirected) as G:
             params = {
                 k: v
                 for k, v in kwargs.items()
@@ -349,6 +356,7 @@ def hits(self, **kwargs):
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.hits(
+            undirected=arguments.get("undirected", False),
             nodes=arguments.get("nodes"),
             nodeIdentifierProperty=arguments.get("nodeIdentifierProperty"),
             hitsIterations=arguments.get("hitsIterations"),
diff --git a/mcp_server/src/mcp_server_neo4j_gds/centrality_algorithm_specs.py b/mcp_server/src/mcp_server_neo4j_gds/centrality_algorithm_specs.py
index 3c71dbf..00bdb96 100644
--- a/mcp_server/src/mcp_server_neo4j_gds/centrality_algorithm_specs.py
+++ b/mcp_server/src/mcp_server_neo4j_gds/centrality_algorithm_specs.py
@@ -60,6 +60,10 @@
                     "Supported values are None, MinMax, Max, Mean, Log, and StdScore. "
                     "To apply scaler-specific configuration, use the Map syntax: {scaler: 'name', ...}.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": [],
         },
@@ -101,6 +105,10 @@
                     "type": "string",
                     "description": "Property of the relationship to use for weighting. If not specified, all relationships are treated equally.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": [],
         },
@@ -142,6 +150,10 @@
                     "type": "string",
                     "description": "Property name to use for identifying nodes (e.g., 'name', 'Name', 'title'). Use get_node_properties_keys to find available properties.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": ["seedSetSize"],
         },
@@ -167,6 +179,10 @@
                     "type": "boolean",
                     "description": "If true, uses the Wasserman-Faust formula for closeness centrality. ",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": [],
         },
@@ -190,6 +206,10 @@
                     "type": "string",
                     "description": "The orientation used to compute node degrees. Supported orientations are NATURAL (for out-degree), REVERSE (for in-degree) and UNDIRECTED (for both in-degree and out-degree) ",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": [],
         },
@@ -256,6 +276,10 @@
                     "Supported values are None, MinMax, Max, Mean, Log, and StdScore. "
                     "To apply scaler-specific configuration, use the Map syntax: {scaler: 'name', ...}.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
         },
     ),
@@ -307,6 +331,10 @@
                         },
                     ],
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": [],
         },
@@ -327,6 +355,10 @@
                     "type": "string",
                     "description": "Property name to use for identifying nodes (e.g., 'name', 'Name', 'title'). Use get_node_properties_keys to find available properties.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": [],
         },
@@ -365,6 +397,10 @@
                     "enum": ["AUTO", "RANGE", "DEGREE"],
                     "description": "The partitioning scheme used to divide the work between threads. Available options are AUTO, RANGE, DEGREE.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": [],
         },
diff --git a/mcp_server/src/mcp_server_neo4j_gds/community_algorithm_handlers.py b/mcp_server/src/mcp_server_neo4j_gds/community_algorithm_handlers.py
index 91ddf1e..1279cd4 100644
--- a/mcp_server/src/mcp_server_neo4j_gds/community_algorithm_handlers.py
+++ b/mcp_server/src/mcp_server_neo4j_gds/community_algorithm_handlers.py
@@ -13,8 +13,8 @@
 
 
 class ConductanceHandler(AlgorithmHandler):
-    def conductance(self, **kwargs):
-        with projected_graph(self.gds) as G:
+    def conductance(self, undirected: bool = False, **kwargs):
+        with projected_graph(self.gds, undirected=undirected) as G:
             logger.info(f"Conductance parameters: {kwargs}")
             conductance = self.gds.conductance.stream(G, **kwargs)
 
@@ -22,6 +22,7 @@ def conductance(self, **kwargs):
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.conductance(
+            undirected=arguments.get("undirected", False),
             communityProperty=arguments.get("communityProperty"),
             relationshipWeightProperty=arguments.get("relationshipWeightProperty"),
         )
@@ -74,8 +75,8 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
 
 
 class K1ColoringHandler(AlgorithmHandler):
-    def k_1_coloring(self, **kwargs):
-        with projected_graph(self.gds) as G:
+    def k_1_coloring(self, undirected: bool = False, **kwargs):
+        with projected_graph(self.gds, undirected=undirected) as G:
             params = {
                 k: v
                 for k, v in kwargs.items()
@@ -94,6 +95,7 @@ def k_1_coloring(self, **kwargs):
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.k_1_coloring(
+            undirected=arguments.get("undirected", False),
             nodeIdentifierProperty=arguments.get("nodeIdentifierProperty"),
             maxIterations=arguments.get("maxIterations"),
             minCommunitySize=arguments.get("minCommunitySize"),
@@ -134,13 +136,13 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
 
 
 class LabelPropagationHandler(AlgorithmHandler):
-    def label_propagation(self, **kwargs):
+    def label_propagation(self, undirected: bool = False, **kwargs):
         # Filter out nodeIdentifierProperty as it's not a GDS algorithm parameter
         gds_kwargs = {
             k: v for k, v in kwargs.items() if k not in ["nodeIdentifierProperty"]
         }
 
-        with projected_graph(self.gds) as G:
+        with projected_graph(self.gds, undirected=undirected) as G:
             logger.info(f"Label Propagation parameters: {gds_kwargs}")
             label_propagation_result = self.gds.labelPropagation.stream(G, **gds_kwargs)
 
@@ -154,6 +156,7 @@ def label_propagation(self, **kwargs):
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.label_propagation(
+            undirected=arguments.get("undirected", False),
             maxIterations=arguments.get("maxIterations"),
             nodeWeightProperty=arguments.get("nodeWeightProperty"),
             relationshipWeightProperty=arguments.get("relationshipWeightProperty"),
@@ -239,13 +242,13 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
 
 
 class LouvainHandler(AlgorithmHandler):
-    def louvain(self, **kwargs):
+    def louvain(self, undirected: bool = False, **kwargs):
         # Filter out nodeIdentifierProperty as it's not a GDS algorithm parameter
         gds_kwargs = {
             k: v for k, v in kwargs.items() if k not in ["nodeIdentifierProperty"]
         }
 
-        with projected_graph(self.gds) as G:
+        with projected_graph(self.gds, undirected=undirected) as G:
             logger.info(f"Louvain parameters: {gds_kwargs}")
             louvain_result = self.gds.louvain.stream(G, **gds_kwargs)
 
@@ -257,6 +260,7 @@ def louvain(self, **kwargs):
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.louvain(
+            undirected=arguments.get("undirected", False),
             relationshipWeightProperty=arguments.get("relationshipWeightProperty"),
             seedProperty=arguments.get("seedProperty"),
             maxLevels=arguments.get("maxLevels"),
@@ -272,8 +276,8 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
 
 
 class ModularityMetricHandler(AlgorithmHandler):
-    def modularity_metric(self, **kwargs):
-        with projected_graph(self.gds) as G:
+    def modularity_metric(self, undirected: bool = False, **kwargs):
+        with projected_graph(self.gds, undirected=undirected) as G:
             logger.info(f"Modularity Metric parameters: {kwargs}")
             modularity_metric_result = self.gds.modularity.stream(G, **kwargs)
 
@@ -281,19 +285,20 @@ def modularity_metric(self, **kwargs):
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.modularity_metric(
+            undirected=arguments.get("undirected", False),
             communityProperty=arguments.get("communityProperty"),
             relationshipWeightProperty=arguments.get("relationshipWeightProperty"),
         )
 
 
 class ModularityOptimizationHandler(AlgorithmHandler):
-    def modularity_optimization(self, **kwargs):
+    def modularity_optimization(self, undirected: bool = False, **kwargs):
         # Filter out nodeIdentifierProperty as it's not a GDS algorithm parameter
         gds_kwargs = {
             k: v for k, v in kwargs.items() if k not in ["nodeIdentifierProperty"]
         }
 
-        with projected_graph(self.gds) as G:
+        with projected_graph(self.gds, undirected=undirected) as G:
             logger.info(f"Modularity Optimization parameters: {gds_kwargs}")
             modularity_optimization_result = self.gds.modularityOptimization.stream(
                 G, **gds_kwargs
@@ -309,6 +314,7 @@ def modularity_optimization(self, **kwargs):
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.modularity_optimization(
+            undirected=arguments.get("undirected", False),
             maxIterations=arguments.get("maxIterations"),
             tolerance=arguments.get("tolerance"),
             seedProperty=arguments.get("seedProperty"),
@@ -413,13 +419,13 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
 
 
 class ApproximateMaximumKCutHandler(AlgorithmHandler):
-    def approximate_maximum_k_cut(self, **kwargs):
+    def approximate_maximum_k_cut(self, undirected: bool = False, **kwargs):
         # Filter out nodeIdentifierProperty as it's not a GDS algorithm parameter
         gds_kwargs = {
             k: v for k, v in kwargs.items() if k not in ["nodeIdentifierProperty"]
         }
 
-        with projected_graph(self.gds) as G:
+        with projected_graph(self.gds, undirected=undirected) as G:
             logger.info(f"Approximate Maximum K Cut parameters: {gds_kwargs}")
             approximate_maximum_k_cut_result = self.gds.maxkcut.stream(G, **gds_kwargs)
 
@@ -433,6 +439,7 @@ def approximate_maximum_k_cut(self, **kwargs):
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.approximate_maximum_k_cut(
+            undirected=arguments.get("undirected", False),
             k=arguments.get("k"),
             iterations=arguments.get("iterations"),
             vnsMaxNeighborhoodOrder=arguments.get("vnsMaxNeighborhoodOrder"),
@@ -443,13 +450,13 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
 
 
 class SpeakerListenerLabelPropagationHandler(AlgorithmHandler):
-    def speaker_listener_label_propagation(self, **kwargs):
+    def speaker_listener_label_propagation(self, undirected: bool = False, **kwargs):
         # Filter out nodeIdentifierProperty as it's not a GDS algorithm parameter
         gds_kwargs = {
             k: v for k, v in kwargs.items() if k not in ["nodeIdentifierProperty"]
         }
 
-        with projected_graph(self.gds) as G:
+        with projected_graph(self.gds, undirected=undirected) as G:
             logger.info(f"Speaker Listener Label Propagation parameters: {gds_kwargs}")
             speaker_listener_label_propagation_result = self.gds.sllpa.stream(
                 G, **gds_kwargs
@@ -467,6 +474,7 @@ def speaker_listener_label_propagation(self, **kwargs):
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.speaker_listener_label_propagation(
+            undirected=arguments.get("undirected", False),
             maxIterations=arguments.get("maxIterations"),
             minAssociationStrength=arguments.get("minAssociationStrength"),
             partitioning=arguments.get("partitioning"),
diff --git a/mcp_server/src/mcp_server_neo4j_gds/community_algorithm_specs.py b/mcp_server/src/mcp_server_neo4j_gds/community_algorithm_specs.py
index 5f74194..328e2ca 100644
--- a/mcp_server/src/mcp_server_neo4j_gds/community_algorithm_specs.py
+++ b/mcp_server/src/mcp_server_neo4j_gds/community_algorithm_specs.py
@@ -17,6 +17,10 @@
                     "description": "The relationship property that holds the weight of the relationships. "
                     "If not provided, all relationships are considered to have a weight of 1.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": ["communityProperty"],
         },
@@ -105,6 +109,10 @@
                     "type": "integer",
                     "description": "Only nodes inside communities larger or equal the given value are returned.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": [],
         },
@@ -202,6 +210,10 @@
                     "type": "string",
                     "description": "The name of a node property to use as node identifier in the result. If provided, the result will include a 'nodeName' column with values from this property.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
         },
     ),
@@ -325,6 +337,10 @@
                     "type": "string",
                     "description": "The name of a node property to use as node identifier in the result. If provided, the result will include a 'nodeName' column with values from this property.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
         },
     ),
@@ -344,6 +360,10 @@
                     "type": "string",
                     "description": "Name of the relationship property to use as weights. If unspecified, the algorithm runs unweighted.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": ["communityProperty"],
         },
@@ -385,6 +405,10 @@
                     "type": "string",
                     "description": "The name of a node property to use as node identifier in the result. If provided, the result will include a 'nodeName' column with values from this property.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
         },
     ),
@@ -510,6 +534,10 @@
                     "type": "string",
                     "description": "The name of a node property to use as node identifier in the result. If provided, the result will include a 'nodeName' column with values from this property.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
         },
     ),
@@ -539,6 +567,10 @@
                     "type": "string",
                     "description": "The name of a node property to use as node identifier in the result. If provided, the result will include a 'nodeName' column with values from this property.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
         },
     ),
diff --git a/mcp_server/src/mcp_server_neo4j_gds/path_algorithm_handlers.py b/mcp_server/src/mcp_server_neo4j_gds/path_algorithm_handlers.py
index 044f760..56c70d0 100644
--- a/mcp_server/src/mcp_server_neo4j_gds/path_algorithm_handlers.py
+++ b/mcp_server/src/mcp_server_neo4j_gds/path_algorithm_handlers.py
@@ -10,7 +10,12 @@
 
 class DijkstraShortestPathHandler(AlgorithmHandler):
     def find_shortest_path(
-        self, start_node: str, end_node: str, node_identifier_property: str, **kwargs
+        self,
+        start_node: str,
+        end_node: str,
+        node_identifier_property: str,
+        undirected: bool = False,
+        **kwargs,
     ):
         query = f"""
         MATCH (start)
@@ -30,7 +35,7 @@ def find_shortest_path(
         start_node_id = int(df["start_id"].iloc[0])
         end_node_id = int(df["end_id"].iloc[0])
 
-        with projected_graph(self.gds) as G:
+        with projected_graph(self.gds, undirected=undirected) as G:
             # If any optional parameter is not None, use that parameter
             args = locals()
             params = {k: v for k, v in kwargs.items() if v is not None}
@@ -72,13 +77,18 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
             arguments.get("start_node"),
             arguments.get("end_node"),
             arguments.get("nodeIdentifierProperty"),
+            undirected=arguments.get("undirected", False),
             relationshipWeightProperty=arguments.get("relationship_property"),
         )
 
 
 class DeltaSteppingShortestPathHandler(AlgorithmHandler):
     def delta_stepping_shortest_path(
-        self, source_node: str, node_identifier_property: str, **kwargs
+        self,
+        source_node: str,
+        node_identifier_property: str,
+        undirected: bool = False,
+        **kwargs,
     ):
         query = f"""
         MATCH (source)
@@ -93,7 +103,7 @@ def delta_stepping_shortest_path(
 
         source_node_id = int(df["source_id"].iloc[0])
 
-        with projected_graph(self.gds) as G:
+        with projected_graph(self.gds, undirected=undirected) as G:
             # If any optional parameter is not None, use that parameter
             params = {k: v for k, v in kwargs.items() if v is not None}
             logger.info(f"Delta-Stepping shortest path parameters: {params}")
@@ -153,6 +163,7 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.delta_stepping_shortest_path(
             arguments.get("sourceNode"),
             arguments.get("nodeIdentifierProperty"),
+            undirected=arguments.get("undirected", False),
             delta=arguments.get("delta"),
             relationshipWeightProperty=arguments.get("relationshipWeightProperty"),
         )
@@ -160,7 +171,11 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
 
 class DijkstraSingleSourceShortestPathHandler(AlgorithmHandler):
     def dijkstra_single_source_shortest_path(
-        self, source_node: str, node_identifier_property: str, **kwargs
+        self,
+        source_node: str,
+        node_identifier_property: str,
+        undirected: bool = False,
+        **kwargs,
     ):
         query = f"""
         MATCH (source)
@@ -175,7 +190,7 @@ def dijkstra_single_source_shortest_path(
 
         source_node_id = int(df["source_id"].iloc[0])
 
-        with projected_graph(self.gds) as G:
+        with projected_graph(self.gds, undirected=undirected) as G:
             # If any optional parameter is not None, use that parameter
             params = {k: v for k, v in kwargs.items() if v is not None}
             logger.info(f"Dijkstra single-source shortest path parameters: {params}")
@@ -234,6 +249,7 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.dijkstra_single_source_shortest_path(
             arguments.get("sourceNode"),
             arguments.get("nodeIdentifierProperty"),
+            undirected=arguments.get("undirected", False),
             relationshipWeightProperty=arguments.get("relationshipWeightProperty"),
         )
 
@@ -244,6 +260,7 @@ def a_star_shortest_path(
         source_node: str,
         target_node: str,
         node_identifier_property: str,
+        undirected: bool = False,
         **kwargs,
     ):
         query = f"""
@@ -264,7 +281,7 @@ def a_star_shortest_path(
         source_node_id = int(df["source_id"].iloc[0])
         target_node_id = int(df["target_id"].iloc[0])
 
-        with projected_graph(self.gds) as G:
+        with projected_graph(self.gds, undirected=undirected) as G:
             # If any optional parameter is not None, use that parameter
             params = {k: v for k, v in kwargs.items() if v is not None}
             logger.info(f"A* shortest path parameters: {params}")
@@ -305,6 +322,7 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
             arguments.get("sourceNode"),
             arguments.get("targetNode"),
             arguments.get("nodeIdentifierProperty"),
+            undirected=arguments.get("undirected", False),
             latitudeProperty=arguments.get("latitudeProperty"),
             longitudeProperty=arguments.get("longitudeProperty"),
             relationshipWeightProperty=arguments.get("relationshipWeightProperty"),
@@ -317,6 +335,7 @@ def yens_shortest_paths(
         source_node: str,
         target_node: str,
         node_identifier_property: str,
+        undirected: bool = False,
         **kwargs,
     ):
         query = f"""
@@ -337,7 +356,7 @@ def yens_shortest_paths(
         source_node_id = int(df["source_id"].iloc[0])
         target_node_id = int(df["target_id"].iloc[0])
 
-        with projected_graph(self.gds) as G:
+        with projected_graph(self.gds, undirected=undirected) as G:
             # If any optional parameter is not None, use that parameter
             params = {k: v for k, v in kwargs.items() if v is not None}
             logger.info(f"Yen's shortest paths parameters: {params}")
@@ -394,6 +413,7 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
             arguments.get("sourceNode"),
             arguments.get("targetNode"),
             arguments.get("nodeIdentifierProperty"),
+            undirected=arguments.get("undirected", False),
             k=arguments.get("k"),
             relationshipWeightProperty=arguments.get("relationshipWeightProperty"),
         )
@@ -481,6 +501,7 @@ def minimum_directed_steiner_tree(
         source_node: str,
         target_nodes: list,
         node_identifier_property: str,
+        undirected: bool = False,
         **kwargs,
     ):
         # Find source node ID
@@ -531,7 +552,7 @@ def minimum_directed_steiner_tree(
         if not target_node_ids:
             return {"found": False, "message": "No target nodes found"}
 
-        with projected_graph(self.gds) as G:
+        with projected_graph(self.gds, undirected=undirected) as G:
             # If any optional parameter is not None, use that parameter
             params = {k: v for k, v in kwargs.items() if v is not None}
             logger.info(f"Minimum Directed Steiner Tree parameters: {params}")
@@ -587,6 +608,7 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
             arguments.get("sourceNode"),
             arguments.get("targetNodes"),
             arguments.get("nodeIdentifierProperty"),
+            undirected=arguments.get("undirected", False),
             relationshipWeightProperty=arguments.get("relationshipWeightProperty"),
             delta=arguments.get("delta"),
             applyRerouting=arguments.get("applyRerouting"),
@@ -652,8 +674,8 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
 
 
 class AllPairsShortestPathsHandler(AlgorithmHandler):
-    def all_pairs_shortest_paths(self, **kwargs):
-        with projected_graph(self.gds) as G:
+    def all_pairs_shortest_paths(self, undirected: bool = False, **kwargs):
+        with projected_graph(self.gds, undirected=undirected) as G:
             # If any optional parameter is not None, use that parameter
             params = {k: v for k, v in kwargs.items() if v is not None}
             logger.info(f"All Pairs Shortest Paths parameters: {params}")
@@ -693,12 +715,13 @@ def all_pairs_shortest_paths(self, **kwargs):
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.all_pairs_shortest_paths(
-            relationshipWeightProperty=arguments.get("relationshipWeightProperty")
+            undirected=arguments.get("undirected", False),
+            relationshipWeightProperty=arguments.get("relationshipWeightProperty"),
         )
 
 
 class RandomWalkHandler(AlgorithmHandler):
-    def random_walk(self, **kwargs):
+    def random_walk(self, undirected: bool = False, **kwargs):
         # Process source nodes if provided
         source_node_ids = []
         if "sourceNodes" in kwargs and kwargs["sourceNodes"]:
@@ -723,7 +746,7 @@ def random_walk(self, **kwargs):
                 if not source_df.empty:
                     source_node_ids.append(int(source_df["source_id"].iloc[0]))
 
-        with projected_graph(self.gds) as G:
+        with projected_graph(self.gds, undirected=undirected) as G:
             # Prepare parameters for the random walk algorithm, excluding our internal parameters
             params = {
                 k: v
@@ -770,6 +793,7 @@ def random_walk(self, **kwargs):
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.random_walk(
+            undirected=arguments.get("undirected", False),
             sourceNodes=arguments.get("sourceNodes"),
             nodeIdentifierProperty=arguments.get("nodeIdentifierProperty"),
             walkLength=arguments.get("walkLength"),
@@ -783,7 +807,11 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
 
 class BreadthFirstSearchHandler(AlgorithmHandler):
     def breadth_first_search(
-        self, source_node: str, node_identifier_property: str, **kwargs
+        self,
+        source_node: str,
+        node_identifier_property: str,
+        undirected: bool = False,
+        **kwargs,
     ):
         # Find source node ID
         source_query = f"""
@@ -818,7 +846,7 @@ def breadth_first_search(
                 if not target_df.empty:
                     target_node_ids.append(int(target_df["target_id"].iloc[0]))
 
-        with projected_graph(self.gds) as G:
+        with projected_graph(self.gds, undirected=undirected) as G:
             # Prepare parameters for the BFS algorithm, excluding our internal parameters
             params = {
                 k: v
@@ -873,6 +901,7 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.breadth_first_search(
             arguments.get("sourceNode"),
             arguments.get("nodeIdentifierProperty"),
+            undirected=arguments.get("undirected", False),
             targetNodes=arguments.get("targetNodes"),
             maxDepth=arguments.get("maxDepth"),
         )
@@ -880,7 +909,11 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
 
 class DepthFirstSearchHandler(AlgorithmHandler):
     def depth_first_search(
-        self, source_node: str, node_identifier_property: str, **kwargs
+        self,
+        source_node: str,
+        node_identifier_property: str,
+        undirected: bool = False,
+        **kwargs,
     ):
         # Find source node ID
         source_query = f"""
@@ -915,7 +948,7 @@ def depth_first_search(
                 if not target_df.empty:
                     target_node_ids.append(int(target_df["target_id"].iloc[0]))
 
-        with projected_graph(self.gds) as G:
+        with projected_graph(self.gds, undirected=undirected) as G:
             # Prepare parameters for the DFS algorithm, excluding our internal parameters
             params = {
                 k: v
@@ -970,6 +1003,7 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.depth_first_search(
             arguments.get("sourceNode"),
             arguments.get("nodeIdentifierProperty"),
+            undirected=arguments.get("undirected", False),
             targetNodes=arguments.get("targetNodes"),
             maxDepth=arguments.get("maxDepth"),
         )
@@ -977,7 +1011,11 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
 
 class BellmanFordSingleSourceShortestPathHandler(AlgorithmHandler):
     def bellman_ford_single_source_shortest_path(
-        self, source_node: str, node_identifier_property: str, **kwargs
+        self,
+        source_node: str,
+        node_identifier_property: str,
+        undirected: bool = False,
+        **kwargs,
     ):
         # Find source node ID
         source_query = f"""
@@ -995,7 +1033,7 @@ def bellman_ford_single_source_shortest_path(
 
         source_node_id = int(source_df["source_id"].iloc[0])
 
-        with projected_graph(self.gds) as G:
+        with projected_graph(self.gds, undirected=undirected) as G:
             # Prepare parameters for the Bellman-Ford algorithm, excluding our internal parameters
             params = {
                 k: v
@@ -1060,6 +1098,7 @@ def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.bellman_ford_single_source_shortest_path(
             arguments.get("sourceNode"),
             arguments.get("nodeIdentifierProperty"),
+            undirected=arguments.get("undirected", False),
             relationshipWeightProperty=arguments.get("relationshipWeightProperty"),
         )
 
diff --git a/mcp_server/src/mcp_server_neo4j_gds/path_algorithm_specs.py b/mcp_server/src/mcp_server_neo4j_gds/path_algorithm_specs.py
index bf0e669..20a180a 100644
--- a/mcp_server/src/mcp_server_neo4j_gds/path_algorithm_specs.py
+++ b/mcp_server/src/mcp_server_neo4j_gds/path_algorithm_specs.py
@@ -23,6 +23,10 @@
                     "type": "string",
                     "description": "Property of the relationship to use for path finding",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": ["start_node", "end_node", "nodeIdentifierProperty"],
         },
@@ -53,6 +57,10 @@
                     "type": "string",
                     "description": "Name of the relationship property to use as weights. If unspecified, the algorithm runs unweighted.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": ["sourceNode", "nodeIdentifierProperty"],
         },
@@ -78,6 +86,10 @@
                     "type": "string",
                     "description": "Name of the relationship property to use as weights. If unspecified, the algorithm runs unweighted.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": ["sourceNode", "nodeIdentifierProperty"],
         },
@@ -121,6 +133,10 @@
                     "type": "string",
                     "description": "Name of the relationship property to use as weights. If unspecified, the algorithm runs unweighted.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": [
                 "sourceNode",
@@ -166,6 +182,10 @@
                     "type": "string",
                     "description": "Name of the relationship property to use as weights. If unspecified, the algorithm runs unweighted.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": ["sourceNode", "targetNode", "nodeIdentifierProperty"],
         },
@@ -244,6 +264,10 @@
                     "type": "boolean",
                     "description": "If specified, the algorithm will try to improve the outcome via an additional post-processing heuristic.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": ["sourceNode", "targetNodes", "nodeIdentifierProperty"],
         },
@@ -289,7 +313,11 @@
                 "relationshipWeightProperty": {
                     "type": "string",
                     "description": "Name of the relationship property to use as weights. If unspecified, the algorithm runs unweighted.",
-                }
+                },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": [],
         },
@@ -343,6 +371,10 @@
                     "type": "integer",
                     "description": "The number of random walks to complete before starting training.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": [],
         },
@@ -374,6 +406,10 @@
                     "type": "integer",
                     "description": "The maximum distance from the source node at which nodes are visited.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": ["sourceNode", "nodeIdentifierProperty"],
         },
@@ -405,6 +441,10 @@
                     "type": "integer",
                     "description": "The maximum distance from the source node at which nodes are visited.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": ["sourceNode", "nodeIdentifierProperty"],
         },
@@ -437,6 +477,10 @@
                     "type": "string",
                     "description": "Name of the relationship property to use as weights. If unspecified, the algorithm runs unweighted.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
             "required": ["sourceNode", "nodeIdentifierProperty"],
         },
diff --git a/mcp_server/src/mcp_server_neo4j_gds/similarity_algorithm_handlers.py b/mcp_server/src/mcp_server_neo4j_gds/similarity_algorithm_handlers.py
index fbb2d9a..e11a860 100644
--- a/mcp_server/src/mcp_server_neo4j_gds/similarity_algorithm_handlers.py
+++ b/mcp_server/src/mcp_server_neo4j_gds/similarity_algorithm_handlers.py
@@ -12,8 +12,8 @@
 
 
 class NodeSimilarityHandler(AlgorithmHandler):
-    def node_similarity(self, **kwargs):
-        with projected_graph(self.gds) as G:
+    def node_similarity(self, undirected: bool = False, **kwargs):
+        with projected_graph(self.gds, undirected=undirected) as G:
             params = {
                 k: v
                 for k, v in kwargs.items()
@@ -67,6 +67,7 @@ def node_similarity(self, **kwargs):
 
     def execute(self, arguments: Dict[str, Any]) -> Any:
         return self.node_similarity(
+            undirected=arguments.get("undirected", False),
             nodeIdentifierProperty=arguments.get("nodeIdentifierProperty"),
             sourceNodeFilter=arguments.get("sourceNodeFilter"),
             targetNodeFilter=arguments.get("targetNodeFilter"),
diff --git a/mcp_server/src/mcp_server_neo4j_gds/similarity_algorithm_specs.py b/mcp_server/src/mcp_server_neo4j_gds/similarity_algorithm_specs.py
index 446e382..3a45a4d 100644
--- a/mcp_server/src/mcp_server_neo4j_gds/similarity_algorithm_specs.py
+++ b/mcp_server/src/mcp_server_neo4j_gds/similarity_algorithm_specs.py
@@ -64,6 +64,10 @@
                     "type": "string",
                     "description": "Property name to use for identifying nodes (e.g., 'name', 'Name', 'title'). Use get_node_properties_keys to find available properties.",
                 },
+                "undirected": {
+                    "type": "boolean",
+                    "description": "Whether to treat the graph as undirected or not. Default is false (directed).",
+                },
             },
         },
     ),