diff --git a/examples/lmcache/disagg_prefill_lmcache_v0.py b/examples/lmcache/disagg_prefill_lmcache_v0.py
index 7da6fb7aaa23..66cc94185230 100644
--- a/examples/lmcache/disagg_prefill_lmcache_v0.py
+++ b/examples/lmcache/disagg_prefill_lmcache_v0.py
@@ -49,9 +49,10 @@ def run_prefill(prefill_done, prompts):
 
     sampling_params = SamplingParams(temperature=0, top_p=0.95, max_tokens=1)
 
-    ktc = KVTransferConfig.from_cli(
-        '{"kv_connector":"LMCacheConnector","kv_role":"kv_producer","kv_rank":0,"kv_parallel_size":2}'
-    )
+    ktc = KVTransferConfig(kv_connector="LMCacheConnector",
+                           kv_role="kv_producer",
+                           kv_rank=0,
+                           kv_parallel_size=2)
     # Set GPU memory utilization to 0.8 for an A40 GPU with 40GB
     # memory. Reduce the value if your GPU has less memory.
     llm = LLM(model="mistralai/Mistral-7B-Instruct-v0.2",
@@ -78,9 +79,10 @@ def run_decode(prefill_done, prompts, timeout=1):
 
     sampling_params = SamplingParams(temperature=0, top_p=0.95, max_tokens=10)
 
-    ktc = KVTransferConfig.from_cli(
-        '{"kv_connector":"LMCacheConnector","kv_role":"kv_consumer","kv_rank":1,"kv_parallel_size":2}'
-    )
+    ktc = KVTransferConfig(kv_connector="LMCacheConnector",
+                           kv_role="kv_consumer",
+                           kv_rank=1,
+                           kv_parallel_size=2)
     # Set GPU memory utilization to 0.8 for an A40 GPU with 40GB
     # of memory. Reduce the value if your GPU has less memory.
     llm = LLM(model="mistralai/Mistral-7B-Instruct-v0.2",
diff --git a/examples/lmcache/kv_cache_sharing_lmcache_v1.py b/examples/lmcache/kv_cache_sharing_lmcache_v1.py
index af1b4351dd54..7748f8ca6133 100644
--- a/examples/lmcache/kv_cache_sharing_lmcache_v1.py
+++ b/examples/lmcache/kv_cache_sharing_lmcache_v1.py
@@ -49,8 +49,8 @@ def run_store(store_done, prompts):
 
     sampling_params = SamplingParams(temperature=0, top_p=0.95, max_tokens=10)
 
-    ktc = KVTransferConfig.from_cli(
-        '{"kv_connector":"LMCacheConnectorV1", "kv_role":"kv_both"}')
+    ktc = KVTransferConfig(kv_connector="LMCacheConnectorV1",
+                           kv_role="kv_both")
     # Set GPU memory utilization to 0.8 for an A40 GPU with 40GB
     # memory. Reduce the value if your GPU has less memory.
     llm = LLM(model="mistralai/Mistral-7B-Instruct-v0.2",
@@ -76,8 +76,8 @@ def run_retrieve(store_done, prompts, timeout=1):
 
     sampling_params = SamplingParams(temperature=0, top_p=0.95, max_tokens=10)
 
-    ktc = KVTransferConfig.from_cli(
-        '{"kv_connector":"LMCacheConnectorV1", "kv_role":"kv_both"}')
+    ktc = KVTransferConfig(kv_connector="LMCacheConnectorV1",
+                           kv_role="kv_both")
     # Set GPU memory utilization to 0.8 for an A40 GPU with 40GB
     # of memory. Reduce the value if your GPU has less memory.
     llm = LLM(model="mistralai/Mistral-7B-Instruct-v0.2",
diff --git a/examples/offline_inference/disaggregated-prefill-v1/decode_example.py b/examples/offline_inference/disaggregated-prefill-v1/decode_example.py
index 66efbc0c9dee..11918f72feec 100644
--- a/examples/offline_inference/disaggregated-prefill-v1/decode_example.py
+++ b/examples/offline_inference/disaggregated-prefill-v1/decode_example.py
@@ -16,16 +16,17 @@ except FileNotFoundError:
 
 sampling_params = SamplingParams(temperature=0, top_p=0.95, max_tokens=10)
 
-llm = LLM(
-    model="meta-llama/Llama-3.2-1B-Instruct",
-    enforce_eager=True,
-    gpu_memory_utilization=0.8,
-    max_num_batched_tokens=64,
-    max_num_seqs=16,
-    kv_transfer_config=KVTransferConfig.from_cli(
-        '{"kv_connector":"SharedStorageConnector","kv_role":"kv_both",'
-        '"kv_connector_extra_config": {"shared_storage_path": "local_storage"}}'
-    ))  #, max_model_len=2048, max_num_batched_tokens=2048)
+llm = LLM(model="meta-llama/Llama-3.2-1B-Instruct",
+          enforce_eager=True,
+          gpu_memory_utilization=0.8,
+          max_num_batched_tokens=64,
+          max_num_seqs=16,
+          kv_transfer_config=KVTransferConfig(
+              kv_connector="SharedStorageConnector",
+              kv_role="kv_both",
+              kv_connector_extra_config={
+                  "shared_storage_path": "local_storage"
+              }))  #, max_model_len=2048, max_num_batched_tokens=2048)
 
 # 1ST generation (prefill instance)
 outputs = llm.generate(prompts, sampling_params)
diff --git a/examples/offline_inference/disaggregated-prefill-v1/prefill_example.py b/examples/offline_inference/disaggregated-prefill-v1/prefill_example.py
index f7cbf6557d54..798128301e0f 100644
--- a/examples/offline_inference/disaggregated-prefill-v1/prefill_example.py
+++ b/examples/offline_inference/disaggregated-prefill-v1/prefill_example.py
@@ -17,11 +17,12 @@ sampling_params = SamplingParams(temperature=0, top_p=0.95, max_tokens=1)
 llm = LLM(model="meta-llama/Llama-3.2-1B-Instruct",
           enforce_eager=True,
           gpu_memory_utilization=0.8,
-          kv_transfer_config=KVTransferConfig.from_cli(
-              '{"kv_connector":"SharedStorageConnector","kv_role":"kv_both", '
-              '"kv_connector_extra_config": '
-              '{"shared_storage_path": "local_storage"}}')
-          )  #, max_model_len=2048, max_num_batched_tokens=2048)
+          kv_transfer_config=KVTransferConfig(
+              kv_connector="SharedStorageConnector",
+              kv_role="kv_both",
+              kv_connector_extra_config={
+                  "shared_storage_path": "local_storage"
+              }))  #, max_model_len=2048, max_num_batched_tokens=2048)
 
 # 1ST generation (prefill instance)
 outputs = llm.generate(
diff --git a/examples/offline_inference/disaggregated_prefill.py b/examples/offline_inference/disaggregated_prefill.py
index d60985146c5c..bb6fdd48f79e 100644
--- a/examples/offline_inference/disaggregated_prefill.py
+++ b/examples/offline_inference/disaggregated_prefill.py
@@ -32,9 +32,10 @@ def run_prefill(prefill_done):
     # This instance is the prefill node (kv_producer, rank 0).
     # The number of parallel instances for KV cache transfer is set to 2,
     # as required for PyNcclConnector.
-    ktc = KVTransferConfig.from_cli(
-        '{"kv_connector":"PyNcclConnector","kv_role":"kv_producer","kv_rank":0,"kv_parallel_size":2}'
-    )
+    ktc = KVTransferConfig(kv_connector="PyNcclConnector",
+                           kv_role="kv_producer",
+                           kv_rank=0,
+                           kv_parallel_size=2)
 
     # Set GPU memory utilization to 0.8 for an A6000 GPU with 40GB
     # memory. You may need to adjust the value to fit your GPU.
@@ -71,9 +72,10 @@ def run_decode(prefill_done):
     # This instance is the decode node (kv_consumer, rank 1).
     # The number of parallel instances for KV cache transfer is set to 2,
     # as required for PyNcclConnector.
-    ktc = KVTransferConfig.from_cli(
-        '{"kv_connector":"PyNcclConnector","kv_role":"kv_consumer","kv_rank":1,"kv_parallel_size":2}'
-    )
+    ktc = KVTransferConfig(kv_connector="PyNcclConnector",
+                           kv_role="kv_consumer",
+                           kv_rank=1,
+                           kv_parallel_size=2)
 
     # Set GPU memory utilization to 0.8 for an A6000 GPU with 40GB
     # memory. You may need to adjust the value to fit your GPU.