diff --git a/README.md b/README.md
index ffb3b09..68aeb8e 100644
--- a/README.md
+++ b/README.md
@@ -3,6 +3,9 @@
 
 ### Installation
 
+This client uses Platform API `4.37.1`. CServe deployments have been retired;
+the CServe SDK helpers and CLI deployment option are no longer available.
+
 To install without cloning, run the following command:
 ```bash
 pip install git+https://github.com/CentML/centml-python-client.git@main
diff --git a/centml/cli/cluster.py b/centml/cli/cluster.py
index be38602..7145737 100644
--- a/centml/cli/cluster.py
+++ b/centml/cli/cluster.py
@@ -21,17 +21,12 @@
     DeploymentType.INFERENCE_V2: "inference",
     DeploymentType.INFERENCE_V3: "inference",
     DeploymentType.COMPUTE_V2: "compute",
-    # For user, they are all cserve.
-    DeploymentType.CSERVE: "cserve",
-    DeploymentType.CSERVE_V2: "cserve",
-    DeploymentType.CSERVE_V3: "cserve",
     DeploymentType.RAG: "rag",
     DeploymentType.JOB: "job",
 }
 # use latest type to for user requests
 depl_name_to_type_map = {
     "inference": DeploymentType.INFERENCE_V3,
-    "cserve": DeploymentType.CSERVE_V3,
     "compute": DeploymentType.COMPUTE_V2,
     "rag": DeploymentType.RAG,
     "job": DeploymentType.JOB,
@@ -78,7 +73,6 @@ def _format_ssh_key(ssh_key):
 def _get_replica_info(deployment):
     """Extract replica information handling V2/V3 field differences"""
     # Check actual deployment object fields rather than depl_type
-    # since unified get_cserve() can return either V2 or V3 objects
     if hasattr(deployment, 'min_replicas'):
         # V3 deployment response object
         return {"min": deployment.min_replicas, "max": deployment.max_replicas}
@@ -218,8 +212,6 @@ def get(type, id):
             deployment = cclient.get_inference(id)  # handles both V2 and V3
         elif depl_type == DeploymentType.COMPUTE_V2:
             deployment = cclient.get_compute(id)
-        elif depl_type in [DeploymentType.CSERVE_V2, DeploymentType.CSERVE_V3]:
-            deployment = cclient.get_cserve(id)  # handles both V2 and V3
         elif depl_type == DeploymentType.JOB:
             deployment = cclient.get_job(id)
         else:
@@ -268,22 +260,6 @@ def get(type, id):
                     disable_numparse=True,
                 )
             )
-        elif depl_type in [DeploymentType.CSERVE_V2, DeploymentType.CSERVE_V3]:
-            replica_info = _get_replica_info(deployment)
-            display_rows = [
-                ("Hugging face model", deployment.recipe.model),
-                (
-                    "Parallelism",
-                    {
-                        "tensor": deployment.recipe.additional_properties.get("tensor_parallel_size", "N/A"),
-                        "pipeline": deployment.recipe.additional_properties.get("pipeline_parallel_size", "N/A"),
-                    },
-                ),
-                ("Replicas", replica_info),
-                ("Max concurrency", deployment.concurrency or "None"),
-            ]
-
-            click.echo(tabulate(display_rows, tablefmt="rounded_outline", disable_numparse=True))
         elif depl_type == DeploymentType.JOB:
             display_rows = [
                 ("Image", deployment.image_url),
diff --git a/centml/sdk/api.py b/centml/sdk/api.py
index d7c4457..b4a4af2 100644
--- a/centml/sdk/api.py
+++ b/centml/sdk/api.py
@@ -25,8 +25,6 @@
 from centml.sdk import auth
 from centml.sdk.config import settings
 
-STATUS_V3_DEPLOYMENT_TYPES = {DeploymentType.INFERENCE_V3, DeploymentType.CSERVE_V3}
-
 DEFAULT_LOG_PAGE_LINES = 100  # server-side default for max_lines
 MAX_LOG_PAGE_LINES = 5000  # server-side ceiling for max_lines
 # fetch_logs asks for less than the server's own default because the common call tails a
@@ -139,23 +137,6 @@ def get_compute(self, id):
     def get_job(self, id):
         return self._api.get_job_deployment_deployments_job_deployment_id_get(id)
 
-    def get_cserve(self, id):
-        """Get CServe deployment details - automatically handles both V2 and V3 deployments"""
-        # Try V3 first (recommended), fallback to V2 if deployment is V2
-        try:
-            return self._api.get_cserve_v3_deployment_deployments_cserve_v3_deployment_id_get(id)
-        except ApiException as e:
-            # If V3 fails with 404 or similar, try V2
-            if e.status in [404, 400]:  # Deployment might be V2 or endpoint not found
-                try:
-                    return self._api.get_cserve_v2_deployment_deployments_cserve_v2_deployment_id_get(id)
-                except ApiException as v2_error:
-                    # If both fail, raise the original V3 error as it's more likely to be the real issue
-                    raise e from v2_error
-            else:
-                # For other errors (auth, network, etc.), raise immediately
-                raise
-
     def get_dynamo(self, id):
         return self._api.get_dynamo_deployment_deployments_dynamo_deployment_id_get(id)
 
diff --git a/requirements.txt b/requirements.txt
index e4e91d5..5174292 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -5,6 +5,6 @@ pyjwt>=2.8.0
 cryptography==50.0.0
 websockets>=16.0
 pyte>=0.8.0
-platform-api-python-client==4.31.0
+platform-api-python-client==4.37.1
 click>=8.4.1
 typing-extensions>=4.5
diff --git a/setup.py b/setup.py
index 8dd6111..045111e 100644
--- a/setup.py
+++ b/setup.py
@@ -11,7 +11,7 @@
 
 setup(
     name='centml',
-    version='0.7.1',
+    version='0.7.2',
     packages=find_packages(exclude=["tests", "tests.*"]),
     python_requires=">=3.10",
     long_description=open('README.md').read(),