diff --git a/README.md b/README.md
index 86f58da8d2..5db8f26861 100644
--- a/README.md
+++ b/README.md
@@ -61,8 +61,6 @@ The initially supported `Microservices` are described in the below table. More `
 A `Microservices` can be created by using the decorator `register_microservice`. Taking the `embedding microservice` as an example:
 
 ```python
-from langchain_community.embeddings import HuggingFaceHubEmbeddings
-
 from comps import register_microservice, EmbedDoc, ServiceType, TextDoc
 
 
diff --git a/comps/retrievers/deployment/docker_compose/compose.yaml b/comps/retrievers/deployment/docker_compose/compose.yaml
index 708f7819d0..2d82f6570c 100644
--- a/comps/retrievers/deployment/docker_compose/compose.yaml
+++ b/comps/retrievers/deployment/docker_compose/compose.yaml
@@ -3,6 +3,7 @@
 
 include:
   - ../../../third_parties/elasticsearch/deployment/docker_compose/compose.yaml
+  - ../../../third_parties/milvus/deployment/docker_compose/compose.yaml
   - ../../../third_parties/opensearch/deployment/docker_compose/compose.yaml
   - ../../../third_parties/neo4j/deployment/docker_compose/compose.yaml
   - ../../../third_parties/pgvector/deployment/docker_compose/compose.yaml
@@ -49,6 +50,12 @@ services:
     depends_on:
       tei-embedding-serving:
         condition: service_healthy
+      standalone:
+        condition: service_healthy
+      etcd:
+        condition: service_healthy
+      minio:
+        condition: service_healthy
 
   retriever-neo4j:
     extends: retriever
diff --git a/comps/retrievers/src/README_elasticsearch.md b/comps/retrievers/src/README_elasticsearch.md
index 6ff1728995..5905c48b4d 100644
--- a/comps/retrievers/src/README_elasticsearch.md
+++ b/comps/retrievers/src/README_elasticsearch.md
@@ -50,6 +50,7 @@ Please refer to this [readme](../../third_parties/elasticsearch/src/README.md).
 
 ```bash
 export TEI_EMBEDDING_ENDPOINT="http://${your_ip}:6060"
+export HUGGINGFACEHUB_API_TOKEN=${your_hf_api_token}
 export RETRIEVER_COMPONENT_NAME="OPEA_RETRIEVER_ELASTICSEARCH"
 python opea_retrievers_microservice.py
 ```
@@ -64,6 +65,7 @@ export ES_CONNECTION_STRING="http://localhost:9200"
 export INDEX_NAME=${your_index_name}
 export TEI_EMBEDDING_ENDPOINT="http://${your_ip}:6060"
 export RETRIEVER_COMPONENT_NAME="OPEA_RETRIEVER_ELASTICSEARCH"
+export HUGGINGFACEHUB_API_TOKEN=${your_hf_api_token}
 ```
 
 ### 2.2 Build Docker Image
@@ -83,7 +85,8 @@ You can choose one as needed.
 ### 2.3 Run Docker with CLI (Option A)
 
 ```bash
-docker run -d --name="retriever-elasticsearch" -p 7000:7000 --ipc=host -e http_proxy=$http_proxy -e https_proxy=$https_proxy -e ES_CONNECTION_STRING=$ES_CONNECTION_STRING  -e INDEX_NAME=$INDEX_NAME -e TEI_ENDPOINT=$TEI_ENDPOINT opea/retriever:latest
+docker run -d --name="retriever-elasticsearch" -p 7000:7000 --ipc=host -e http_proxy=$http_proxy -e https_proxy=$https_proxy -e ES_CONNECTION_STRING=$ES_CONNECTION_STRING  -e INDEX_NAME=$INDEX_NAME -e TEI_EMBEDDING_ENDPOINT=${TEI_EMBEDDING_ENDPOINT}
+-e HUGGINGFACEHUB_API_TOKEN=${HUGGINGFACEHUB_API_TOKEN} opea/retriever:latest
 ```
 
 ### 2.4 Run Docker with Docker Compose (Option B)
diff --git a/comps/retrievers/src/README_milvus.md b/comps/retrievers/src/README_milvus.md
index 178112c279..d3f94562f9 100644
--- a/comps/retrievers/src/README_milvus.md
+++ b/comps/retrievers/src/README_milvus.md
@@ -22,12 +22,13 @@ export MILVUS_HOST=${your_milvus_host_ip}
 export MILVUS_PORT=19530
 export COLLECTION_NAME=${your_collection_name}
 export TEI_EMBEDDING_ENDPOINT=${your_emdding_endpoint}
+export HUGGINGFACEHUB_API_TOKEN=${your_hf_api_token}
 ```
 
 ### Start Retriever Service
 
 ```bash
-export TEI_EMBEDDING_ENDPOINT="http://${your_ip}:6060"
+export TEI_EMBEDDING_ENDPOINT="http://${your_ip}:${your_embedding_port}"
 export RETRIEVER_COMPONENT_NAME="OPEA_RETRIEVER_MILVUS"
 python opea_retrievers_microservice.py
 ```
@@ -44,7 +45,7 @@ docker build -t opea/retriever:latest --build-arg https_proxy=$https_proxy --bui
 ### Run Docker with CLI (Option A)
 
 ```bash
-docker run -d --name="retriever-milvus-server" -p 7000:7000 --ipc=host -e http_proxy=$http_proxy -e https_proxy=$https_proxy -e TEI_EMBEDDING_ENDPOINT=${your_emdding_endpoint} -e MILVUS_HOST=${your_milvus_host_ip} -e RETRIEVER_COMPONENT_NAME=$RETRIEVER_COMPONENT_NAME opea/retriever:latest
+docker run -d --name="retriever-milvus-server" -p 7000:7000 --ipc=host -e http_proxy=$http_proxy -e https_proxy=$https_proxy -e no_proxy=$no_proxy -e TEI_EMBEDDING_ENDPOINT=${your_emdding_endpoint} -e MILVUS_HOST=${your_milvus_host_ip} -e HUGGINGFACEHUB_API_TOKEN=${your_hf_api_token} -e RETRIEVER_COMPONENT_NAME=$RETRIEVER_COMPONENT_NAME opea/retriever:latest
 ```
 
 ### Run Docker with Docker Compose (Option B)
diff --git a/comps/retrievers/src/README_opensearch.md b/comps/retrievers/src/README_opensearch.md
index 0984553a0c..ee7cf6f39e 100644
--- a/comps/retrievers/src/README_opensearch.md
+++ b/comps/retrievers/src/README_opensearch.md
@@ -44,6 +44,7 @@ Please refer to this [readme](../../third_parties/opensearch/src/README.md).
 ```bash
 export TEI_EMBEDDING_ENDPOINT="http://${your_ip}:6060"
 export RETRIEVER_COMPONENT_NAME="OPEA_RETRIEVER_OPENSEARCH"
+export HUGGINGFACEHUB_API_TOKEN=${your_hf_token}
 python opea_retrievers_microservice.py
 ```
 
diff --git a/comps/retrievers/src/README_vdms.md b/comps/retrievers/src/README_vdms.md
index 1e699cf291..7bfef451ec 100644
--- a/comps/retrievers/src/README_vdms.md
+++ b/comps/retrievers/src/README_vdms.md
@@ -65,6 +65,7 @@ Please refer to this [readme](../../third_parties/vdms/src/README.md).
 
 ```bash
 export TEI_EMBEDDING_ENDPOINT="http://${your_ip}:6060"
+export HUGGINGFACEHUB_API_TOKEN=${your_hf_api_token}
 export RETRIEVER_COMPONENT_NAME="OPEA_RETRIEVER_VDMS"
 python opea_retrievers_microservice.py
 ```
@@ -77,6 +78,7 @@ python opea_retrievers_microservice.py
 export RETRIEVE_MODEL_ID="BAAI/bge-base-en-v1.5"
 export INDEX_NAME=${your_index_name or collection_name}
 export TEI_EMBEDDING_ENDPOINT="http://${your_ip}:6060"
+export HUGGINGFACEHUB_API_TOKEN=${your_hf_api_token}
 export RETRIEVER_COMPONENT_NAME="OPEA_RETRIEVER_VDMS"
 ```
 
@@ -97,7 +99,7 @@ You can choose one as needed.
 ### 2.3 Run Docker with CLI (Option A)
 
 ```bash
-docker run -d --name="retriever-vdms-server" -p 7000:7000 --ipc=host -e http_proxy=$http_proxy -e https_proxy=$https_proxy -e INDEX_NAME=$INDEX_NAME -e TEI_EMBEDDING_ENDPOINT=$TEI_EMBEDDING_ENDPOINT -e RETRIEVER_COMPONENT_NAME=$RETRIEVER_COMPONENT_NAME opea/retriever:latest
+docker run -d --name="retriever-vdms-server" -p 7000:7000 --ipc=host -e http_proxy=$http_proxy -e https_proxy=$https_proxy -e INDEX_NAME=$INDEX_NAME -e TEI_EMBEDDING_ENDPOINT=$TEI_EMBEDDING_ENDPOINT -e HUGGINGFACEHUB_API_TOKEN=${HUGGINGFACEHUB_API_TOKEN} -e RETRIEVER_COMPONENT_NAME=$RETRIEVER_COMPONENT_NAME opea/retriever:latest
 ```
 
 ### 2.4 Run Docker with Docker Compose (Option B)
diff --git a/comps/retrievers/src/integrations/config.py b/comps/retrievers/src/integrations/config.py
index 3ffced0820..45149c0e4b 100644
--- a/comps/retrievers/src/integrations/config.py
+++ b/comps/retrievers/src/integrations/config.py
@@ -44,8 +44,9 @@ def get_boolean_env_var(var_name, default_value=False):
 # Embedding model
 EMBED_MODEL = os.getenv("EMBED_MODEL", "BAAI/bge-base-en-v1.5")
 LOCAL_EMBEDDING_MODEL = os.getenv("LOCAL_EMBEDDING_MODEL", "maidalun1020/bce-embedding-base_v1")
-TEI_EMBEDDING_ENDPOINT = os.getenv("TEI_EMBEDDING_ENDPOINT")
+TEI_EMBEDDING_ENDPOINT = os.getenv("TEI_EMBEDDING_ENDPOINT", "")
 BRIDGE_TOWER_EMBEDDING = os.getenv("BRIDGE_TOWER_EMBEDDING", False)
+HUGGINGFACEHUB_API_TOKEN = os.getenv("HUGGINGFACEHUB_API_TOKEN", "")
 
 # Directory pathss
 current_file_path = os.path.abspath(__file__)
@@ -122,7 +123,6 @@ def format_redis_conn_from_env():
 COLLECTION_NAME = os.getenv("COLLECTION_NAME", "rag_milvus")
 # TEI configuration
 TEI_EMBEDDING_MODEL = os.environ.get("TEI_EMBEDDING_MODEL", "/home/user/bce-embedding-base_v1")
-TEI_EMBEDDING_ENDPOINT = os.environ.get("TEI_EMBEDDING_ENDPOINT", "")
 os.environ["OPENAI_API_BASE"] = TEI_EMBEDDING_ENDPOINT
 os.environ["OPENAI_API_KEY"] = "Dummy key"
 
diff --git a/comps/retrievers/src/integrations/elasticsearch.py b/comps/retrievers/src/integrations/elasticsearch.py
index 08b358e398..72dc2952f4 100644
--- a/comps/retrievers/src/integrations/elasticsearch.py
+++ b/comps/retrievers/src/integrations/elasticsearch.py
@@ -5,12 +5,13 @@
 import os
 
 from elasticsearch import Elasticsearch
-from langchain_community.embeddings import HuggingFaceBgeEmbeddings, HuggingFaceHubEmbeddings
+from fastapi import HTTPException
+from langchain_community.embeddings import HuggingFaceBgeEmbeddings, HuggingFaceInferenceAPIEmbeddings
 from langchain_elasticsearch import ElasticsearchStore
 
 from comps import CustomLogger, EmbedDoc, OpeaComponent, OpeaComponentRegistry, ServiceType
 
-from .config import EMBED_MODEL, ES_CONNECTION_STRING, ES_INDEX_NAME, TEI_EMBEDDING_ENDPOINT
+from .config import EMBED_MODEL, ES_CONNECTION_STRING, ES_INDEX_NAME, HUGGINGFACEHUB_API_TOKEN, TEI_EMBEDDING_ENDPOINT
 
 logger = CustomLogger("es_retrievers")
 logflag = os.getenv("LOGFLAG", False)
@@ -40,7 +41,22 @@ def _initialize_embedder(self):
             # create embeddings using TEI endpoint service
             if logflag:
                 logger.info(f"[ init embedder ] TEI_EMBEDDING_ENDPOINT:{TEI_EMBEDDING_ENDPOINT}")
-            embeddings = HuggingFaceHubEmbeddings(model=TEI_EMBEDDING_ENDPOINT)
+            if not HUGGINGFACEHUB_API_TOKEN:
+                raise HTTPException(
+                    status_code=400,
+                    detail="You MUST offer the `HUGGINGFACEHUB_API_TOKEN` when using `TEI_EMBEDDING_ENDPOINT`.",
+                )
+            import requests
+
+            response = requests.get(TEI_EMBEDDING_ENDPOINT + "/info")
+            if response.status_code != 200:
+                raise HTTPException(
+                    status_code=400, detail=f"TEI embedding endpoint {TEI_EMBEDDING_ENDPOINT} is not available."
+                )
+            model_id = response.json()["model_id"]
+            embeddings = HuggingFaceInferenceAPIEmbeddings(
+                api_key=HUGGINGFACEHUB_API_TOKEN, model_name=model_id, api_url=TEI_EMBEDDING_ENDPOINT
+            )
         else:
             # create embeddings using local embedding model
             if logflag:
diff --git a/comps/retrievers/src/integrations/milvus.py b/comps/retrievers/src/integrations/milvus.py
index b78c33c024..a8655a29c5 100644
--- a/comps/retrievers/src/integrations/milvus.py
+++ b/comps/retrievers/src/integrations/milvus.py
@@ -3,14 +3,21 @@
 
 
 import os
-from typing import List, Optional
 
-from langchain_community.embeddings import HuggingFaceBgeEmbeddings, HuggingFaceHubEmbeddings
+from fastapi import HTTPException
+from langchain_community.embeddings import HuggingFaceBgeEmbeddings, HuggingFaceInferenceAPIEmbeddings
 from langchain_milvus.vectorstores import Milvus
 
 from comps import CustomLogger, EmbedDoc, OpeaComponent, OpeaComponentRegistry, ServiceType
 
-from .config import COLLECTION_NAME, INDEX_PARAMS, LOCAL_EMBEDDING_MODEL, MILVUS_URI, TEI_EMBEDDING_ENDPOINT
+from .config import (
+    COLLECTION_NAME,
+    HUGGINGFACEHUB_API_TOKEN,
+    INDEX_PARAMS,
+    LOCAL_EMBEDDING_MODEL,
+    MILVUS_URI,
+    TEI_EMBEDDING_ENDPOINT,
+)
 
 logger = CustomLogger("milvus_retrievers")
 logflag = os.getenv("LOGFLAG", False)
@@ -28,7 +35,6 @@ def __init__(self, name: str, description: str, config: dict = None):
         super().__init__(name, ServiceType.RETRIEVER.name.lower(), description, config)
 
         self.embedder = self._initialize_embedder()
-        self.client = self._initialize_client()
         health_status = self.check_health()
         if not health_status:
             logger.error("OpeaMilvusRetriever health check failed.")
@@ -38,7 +44,22 @@ def _initialize_embedder(self):
             # create embeddings using TEI endpoint service
             if logflag:
                 logger.info(f"[ init embedder ] TEI_EMBEDDING_ENDPOINT:{TEI_EMBEDDING_ENDPOINT}")
-            embeddings = HuggingFaceHubEmbeddings(model=TEI_EMBEDDING_ENDPOINT)
+            if not HUGGINGFACEHUB_API_TOKEN:
+                raise HTTPException(
+                    status_code=400,
+                    detail="You MUST offer the `HUGGINGFACEHUB_API_TOKEN` when using `TEI_EMBEDDING_ENDPOINT`.",
+                )
+            import requests
+
+            response = requests.get(TEI_EMBEDDING_ENDPOINT + "/info")
+            if response.status_code != 200:
+                raise HTTPException(
+                    status_code=400, detail=f"TEI embedding endpoint {TEI_EMBEDDING_ENDPOINT} is not available."
+                )
+            model_id = response.json()["model_id"]
+            embeddings = HuggingFaceInferenceAPIEmbeddings(
+                api_key=HUGGINGFACEHUB_API_TOKEN, model_name=model_id, api_url=TEI_EMBEDDING_ENDPOINT
+            )
         else:
             # create embeddings using local embedding model
             if logflag:
@@ -70,7 +91,14 @@ def check_health(self) -> bool:
         if logflag:
             logger.info("[ check health ] start to check health of milvus")
         try:
-            _ = self.client.client.list_collections()
+            client = Milvus(
+                embedding_function=self.embedder,
+                collection_name=COLLECTION_NAME,
+                connection_args={"uri": MILVUS_URI},
+                index_params=INDEX_PARAMS,
+                auto_id=True,
+            )
+            _ = client.client.list_collections()
             if logflag:
                 logger.info("[ check health ] Successfully connected to Milvus!")
             return True
@@ -89,21 +117,29 @@ async def invoke(self, input: EmbedDoc) -> list:
         if logflag:
             logger.info(input)
 
+        my_milvus = Milvus(
+            embedding_function=self.embedder,
+            collection_name=COLLECTION_NAME,
+            connection_args={"uri": MILVUS_URI},
+            index_params=INDEX_PARAMS,
+            auto_id=True,
+        )
+
         if input.search_type == "similarity":
-            search_res = await self.client.asimilarity_search_by_vector(embedding=input.embedding, k=input.k)
+            search_res = await my_milvus.asimilarity_search_by_vector(embedding=input.embedding, k=input.k)
         elif input.search_type == "similarity_distance_threshold":
             if input.distance_threshold is None:
                 raise ValueError("distance_threshold must be provided for " + "similarity_distance_threshold retriever")
-            search_res = await self.client.asimilarity_search_by_vector(
+            search_res = await my_milvus.asimilarity_search_by_vector(
                 embedding=input.embedding, k=input.k, distance_threshold=input.distance_threshold
             )
         elif input.search_type == "similarity_score_threshold":
-            docs_and_similarities = await self.client.asimilarity_search_with_relevance_scores(
+            docs_and_similarities = await my_milvus.asimilarity_search_with_relevance_scores(
                 query=input.text, k=input.k, score_threshold=input.score_threshold
             )
             search_res = [doc for doc, _ in docs_and_similarities]
         elif input.search_type == "mmr":
-            search_res = await self.client.amax_marginal_relevance_search(
+            search_res = await my_milvus.amax_marginal_relevance_search(
                 query=input.text, k=input.k, fetch_k=input.fetch_k, lambda_mult=input.lambda_mult
             )
 
diff --git a/comps/retrievers/src/integrations/opensearch.py b/comps/retrievers/src/integrations/opensearch.py
index 30514d7107..229ee078e2 100644
--- a/comps/retrievers/src/integrations/opensearch.py
+++ b/comps/retrievers/src/integrations/opensearch.py
@@ -6,7 +6,8 @@
 from typing import Callable, List, Union
 
 import numpy as np
-from langchain_community.embeddings import HuggingFaceBgeEmbeddings, HuggingFaceHubEmbeddings
+from fastapi import HTTPException
+from langchain_community.embeddings import HuggingFaceBgeEmbeddings, HuggingFaceInferenceAPIEmbeddings
 from langchain_community.vectorstores import OpenSearchVectorSearch
 from pydantic import conlist
 
@@ -15,6 +16,7 @@
 
 from .config import (
     EMBED_MODEL,
+    HUGGINGFACEHUB_API_TOKEN,
     OPENSEARCH_INDEX_NAME,
     OPENSEARCH_INITIAL_ADMIN_PASSWORD,
     OPENSEARCH_URL,
@@ -49,7 +51,22 @@ def _initialize_embedder(self):
             # create embeddings using TEI endpoint service
             if logflag:
                 logger.info(f"[ init embedder ] TEI_EMBEDDING_ENDPOINT:{TEI_EMBEDDING_ENDPOINT}")
-            embeddings = HuggingFaceHubEmbeddings(model=TEI_EMBEDDING_ENDPOINT)
+            if not HUGGINGFACEHUB_API_TOKEN:
+                raise HTTPException(
+                    status_code=400,
+                    detail="You MUST offer the `HUGGINGFACEHUB_API_TOKEN` when using `TEI_EMBEDDING_ENDPOINT`.",
+                )
+            import requests
+
+            response = requests.get(TEI_EMBEDDING_ENDPOINT + "/info")
+            if response.status_code != 200:
+                raise HTTPException(
+                    status_code=400, detail=f"TEI embedding endpoint {TEI_EMBEDDING_ENDPOINT} is not available."
+                )
+            model_id = response.json()["model_id"]
+            embeddings = HuggingFaceInferenceAPIEmbeddings(
+                api_key=HUGGINGFACEHUB_API_TOKEN, model_name=model_id, api_url=TEI_EMBEDDING_ENDPOINT
+            )
         else:
             # create embeddings using local embedding model
             if logflag:
diff --git a/comps/retrievers/src/integrations/pgvector.py b/comps/retrievers/src/integrations/pgvector.py
index 164a382473..81d913c2d2 100644
--- a/comps/retrievers/src/integrations/pgvector.py
+++ b/comps/retrievers/src/integrations/pgvector.py
@@ -4,12 +4,13 @@
 
 import os
 
-from langchain_community.embeddings import HuggingFaceBgeEmbeddings, HuggingFaceHubEmbeddings
+from fastapi import HTTPException
+from langchain_community.embeddings import HuggingFaceBgeEmbeddings, HuggingFaceInferenceAPIEmbeddings
 from langchain_community.vectorstores import PGVector
 
 from comps import CustomLogger, EmbedDoc, OpeaComponent, OpeaComponentRegistry, ServiceType
 
-from .config import EMBED_MODEL, PG_CONNECTION_STRING, PG_INDEX_NAME, TEI_EMBEDDING_ENDPOINT
+from .config import EMBED_MODEL, HUGGINGFACEHUB_API_TOKEN, PG_CONNECTION_STRING, PG_INDEX_NAME, TEI_EMBEDDING_ENDPOINT
 
 logger = CustomLogger("pgvector_retrievers")
 logflag = os.getenv("LOGFLAG", False)
@@ -39,7 +40,22 @@ def _initialize_embedder(self):
             # create embeddings using TEI endpoint service
             if logflag:
                 logger.info(f"[ init embedder ] TEI_EMBEDDING_ENDPOINT:{TEI_EMBEDDING_ENDPOINT}")
-            embeddings = HuggingFaceHubEmbeddings(model=TEI_EMBEDDING_ENDPOINT)
+            if not HUGGINGFACEHUB_API_TOKEN:
+                raise HTTPException(
+                    status_code=400,
+                    detail="You MUST offer the `HUGGINGFACEHUB_API_TOKEN` when using `TEI_EMBEDDING_ENDPOINT`.",
+                )
+            import requests
+
+            response = requests.get(TEI_EMBEDDING_ENDPOINT + "/info")
+            if response.status_code != 200:
+                raise HTTPException(
+                    status_code=400, detail=f"TEI embedding endpoint {TEI_EMBEDDING_ENDPOINT} is not available."
+                )
+            model_id = response.json()["model_id"]
+            embeddings = HuggingFaceInferenceAPIEmbeddings(
+                api_key=HUGGINGFACEHUB_API_TOKEN, model_name=model_id, api_url=TEI_EMBEDDING_ENDPOINT
+            )
         else:
             # create embeddings using local embedding model
             if logflag:
diff --git a/comps/retrievers/src/integrations/pinecone.py b/comps/retrievers/src/integrations/pinecone.py
index c8c048a42b..a6713d84ed 100644
--- a/comps/retrievers/src/integrations/pinecone.py
+++ b/comps/retrievers/src/integrations/pinecone.py
@@ -5,13 +5,14 @@
 import os
 import time
 
-from langchain_community.embeddings import HuggingFaceBgeEmbeddings, HuggingFaceHubEmbeddings
+from fastapi import HTTPException
+from langchain_community.embeddings import HuggingFaceBgeEmbeddings, HuggingFaceInferenceAPIEmbeddings
 from langchain_pinecone import PineconeVectorStore
 from pinecone import Pinecone, ServerlessSpec
 
 from comps import CustomLogger, EmbedDoc, OpeaComponent, OpeaComponentRegistry, ServiceType
 
-from .config import EMBED_MODEL, PINECONE_API_KEY, PINECONE_INDEX_NAME, TEI_EMBEDDING_ENDPOINT
+from .config import EMBED_MODEL, HUGGINGFACEHUB_API_TOKEN, PINECONE_API_KEY, PINECONE_INDEX_NAME, TEI_EMBEDDING_ENDPOINT
 
 logger = CustomLogger("pinecone_retrievers")
 logflag = os.getenv("LOGFLAG", False)
@@ -41,7 +42,22 @@ def _initialize_embedder(self):
             # create embeddings using TEI endpoint service
             if logflag:
                 logger.info(f"[ init embedder ] TEI_EMBEDDING_ENDPOINT:{TEI_EMBEDDING_ENDPOINT}")
-            embeddings = HuggingFaceHubEmbeddings(model=TEI_EMBEDDING_ENDPOINT)
+            if not HUGGINGFACEHUB_API_TOKEN:
+                raise HTTPException(
+                    status_code=400,
+                    detail="You MUST offer the `HUGGINGFACEHUB_API_TOKEN` when using `TEI_EMBEDDING_ENDPOINT`.",
+                )
+            import requests
+
+            response = requests.get(TEI_EMBEDDING_ENDPOINT + "/info")
+            if response.status_code != 200:
+                raise HTTPException(
+                    status_code=400, detail=f"TEI embedding endpoint {TEI_EMBEDDING_ENDPOINT} is not available."
+                )
+            model_id = response.json()["model_id"]
+            embeddings = HuggingFaceInferenceAPIEmbeddings(
+                api_key=HUGGINGFACEHUB_API_TOKEN, model_name=model_id, api_url=TEI_EMBEDDING_ENDPOINT
+            )
         else:
             # create embeddings using local embedding model
             if logflag:
diff --git a/comps/retrievers/src/integrations/vdms.py b/comps/retrievers/src/integrations/vdms.py
index 5e5b1731fa..9b8d500da6 100644
--- a/comps/retrievers/src/integrations/vdms.py
+++ b/comps/retrievers/src/integrations/vdms.py
@@ -3,9 +3,9 @@
 
 
 import os
-import time
 
-from langchain_community.embeddings import HuggingFaceBgeEmbeddings, HuggingFaceHubEmbeddings
+from fastapi import HTTPException
+from langchain_community.embeddings import HuggingFaceBgeEmbeddings, HuggingFaceInferenceAPIEmbeddings
 from langchain_community.vectorstores.vdms import VDMS, VDMS_Client
 
 from comps import CustomLogger, EmbedDoc, OpeaComponent, OpeaComponentRegistry, ServiceType
@@ -13,6 +13,7 @@
 from .config import (
     DISTANCE_STRATEGY,
     EMBED_MODEL,
+    HUGGINGFACEHUB_API_TOKEN,
     SEARCH_ENGINE,
     TEI_EMBEDDING_ENDPOINT,
     VDMS_HOST,
@@ -52,7 +53,22 @@ def _initialize_embedder(self):
             # create embeddings using TEI endpoint service
             if logflag:
                 logger.info(f"[ init embedder ] TEI_EMBEDDING_ENDPOINT:{TEI_EMBEDDING_ENDPOINT}")
-            embeddings = HuggingFaceHubEmbeddings(model=TEI_EMBEDDING_ENDPOINT)
+            if not HUGGINGFACEHUB_API_TOKEN:
+                raise HTTPException(
+                    status_code=400,
+                    detail="You MUST offer the `HUGGINGFACEHUB_API_TOKEN` when using `TEI_EMBEDDING_ENDPOINT`.",
+                )
+            import requests
+
+            response = requests.get(TEI_EMBEDDING_ENDPOINT + "/info")
+            if response.status_code != 200:
+                raise HTTPException(
+                    status_code=400, detail=f"TEI embedding endpoint {TEI_EMBEDDING_ENDPOINT} is not available."
+                )
+            model_id = response.json()["model_id"]
+            embeddings = HuggingFaceInferenceAPIEmbeddings(
+                api_key=HUGGINGFACEHUB_API_TOKEN, model_name=model_id, api_url=TEI_EMBEDDING_ENDPOINT
+            )
         else:
             # create embeddings using local embedding model
             if logflag:
diff --git a/comps/third_parties/milvus/deployment/docker_compose/compose.yaml b/comps/third_parties/milvus/deployment/docker_compose/compose.yaml
index abc6ae1883..7eef87687c 100644
--- a/comps/third_parties/milvus/deployment/docker_compose/compose.yaml
+++ b/comps/third_parties/milvus/deployment/docker_compose/compose.yaml
@@ -61,7 +61,3 @@ services:
     depends_on:
       - "etcd"
       - "minio"
-
-networks:
-  default:
-    name: milvus
diff --git a/tests/retrievers/test_retrievers_milvus.sh b/tests/retrievers/test_retrievers_milvus.sh
index 05bd69601a..0dc8146f84 100644
--- a/tests/retrievers/test_retrievers_milvus.sh
+++ b/tests/retrievers/test_retrievers_milvus.sh
@@ -13,7 +13,8 @@ echo "TAG=${TAG}"
 WORKPATH=$(dirname "$PWD")
 LOG_PATH="$WORKPATH/tests"
 export host_ip=$(hostname -I | awk '{print $1}')
-service_name="retriever-milvus"
+service_name="retriever-milvus etcd minio standalone"
+retriever_service_name="retriever-milvus"
 
 function build_docker_images() {
     cd $WORKPATH
@@ -27,7 +28,7 @@ function build_docker_images() {
 }
 
 function start_service() {
-    export MINIO_PORT1=11611   # 11600-11699
+    export MINIO_PORT1=11611
     export MINIO_PORT2=11612
     export MILVUS_STANDALONE_PORT=11613
     export TEI_EMBEDDER_PORT=11614
@@ -38,13 +39,6 @@ function start_service() {
     export TEI_EMBEDDING_ENDPOINT="http://${host_ip}:${TEI_EMBEDDER_PORT}"
     export MILVUS_HOST=${host_ip}
 
-    # wget https://raw.githubusercontent.com/milvus-io/milvus/v2.4.9/configs/milvus.yaml
-    # wget https://github.com/milvus-io/milvus/releases/download/v2.4.9/milvus-standalone-docker-compose.yml -O docker-compose.yml
-    # sed '/- \${DOCKER_VOLUME_DIRECTORY:-\.}\/volumes\/milvus:\/var\/lib\/milvus/a \ \ \ \ \ \ - \${DOCKER_VOLUME_DIRECTORY:-\.}\/milvus.yaml:\/milvus\/configs\/milvus.yaml' -i docker-compose.yml
-
-    cd $WORKPATH/comps/third_parties/milvus/deployment/docker_compose/
-    docker compose -f compose.yaml up -d > ${LOG_PATH}/start_services_with_compose_milvus.log
-
     cd $WORKPATH/comps/retrievers/deployment/docker_compose
     docker compose -f compose.yaml up ${service_name} -d > ${LOG_PATH}/start_services_with_compose.log
 
@@ -67,12 +61,12 @@ function validate_microservice() {
             echo "[ retriever ] Content is as expected."
         else
             echo "[ retriever ] Content does not match the expected result: $CONTENT"
-            docker logs ${service_name} >> ${LOG_PATH}/retriever.log
+            docker logs ${retriever_service_name} >> ${LOG_PATH}/retriever.log
             exit 1
         fi
     else
         echo "[ retriever ] HTTP status is not 200. Received status was $HTTP_STATUS"
-        docker logs ${service_name} >> ${LOG_PATH}/retriever.log
+        docker logs ${retriever_service_name} >> ${LOG_PATH}/retriever.log
         exit 1
     fi
 }
diff --git a/tests/retrievers/test_retrievers_neo4j_on_intel_hpu.sh b/tests/retrievers/test_retrievers_neo4j_on_intel_hpu.sh
index a819e2e485..89dde5d4b2 100644
--- a/tests/retrievers/test_retrievers_neo4j_on_intel_hpu.sh
+++ b/tests/retrievers/test_retrievers_neo4j_on_intel_hpu.sh
@@ -39,8 +39,6 @@ function build_docker_images() {
 }
 
 function start_service() {
-    export NEO4J_PORT1=11631
-    export NEO4J_PORT2=11632
     export TEI_EMBEDDER_PORT=11633
     export LLM_ENDPOINT_PORT=11634
     export RETRIEVER_PORT=11635
@@ -52,7 +50,7 @@ function start_service() {
     export LLM_MODEL_ID="meta-llama/Meta-Llama-3.1-8B-Instruct"
     export TGI_LLM_ENDPOINT="http://${host_ip}:${LLM_ENDPOINT_PORT}"
     export DATAPREP_SERVICE_ENDPOINT="http://${host_ip}:6004"
-    export NEO4J_URI="bolt://${host_ip}:${NEO4J_PORT2}"
+    export NEO4J_URI="bolt://${host_ip}:7687"
     export NEO4J_USERNAME="neo4j"
     export NEO4J_PASSWORD="neo4jtest"
     export no_proxy="localhost,127.0.0.1,"${host_ip}
@@ -65,9 +63,10 @@ function start_service() {
     # dataprep neo4j
     # Not testing openai code path since not able to provide key for cicd
     docker run -d --name="test-comps-retrievers-neo4j-llama-index-dataprep" -p 6004:5000 -v ./data:/data --ipc=host -e TGI_LLM_ENDPOINT=$TGI_LLM_ENDPOINT \
-        -e TEI_EMBEDDING_ENDPOINT=$TEI_EMBEDDING_ENDPOINT -e TEI_ENDPOINT=$TEI_EMBEDDING_ENDPOINT -e EMBEDDING_MODEL_ID=$EMBEDDING_MODEL_ID -e LLM_MODEL_ID=$LLM_MODEL_ID -e host_ip=$host_ip -e no_proxy=$no_proxy \
-        -e http_proxy=$http_proxy -e https_proxy=$https_proxy -e NEO4J_URL="bolt://${host_ip}:${NEO4J_PORT2}" -e NEO4J_USERNAME="neo4j" \
-        -e NEO4J_PASSWORD="neo4jtest" -e HF_TOKEN=$HF_TOKEN -e MAX_INPUT_LEN=$MAX_INPUT_TOKENS -e LOGFLAG=True -e DATAPREP_COMPONENT_NAME="OPEA_DATAPREP_NEO4J_LLAMAINDEX" opea/dataprep-neo4j-llamaindex:comps
+        -e TEI_EMBEDDING_ENDPOINT=$TEI_EMBEDDING_ENDPOINT -e EMBEDDING_MODEL_ID=$EMBEDDING_MODEL_ID -e LLM_MODEL_ID=$LLM_MODEL_ID -e host_ip=$host_ip -e no_proxy=$no_proxy \
+        -e http_proxy=$http_proxy -e https_proxy=$https_proxy -e NEO4J_URL="bolt://${host_ip}:7687" -e NEO4J_USERNAME="neo4j" \
+        -e NEO4J_PASSWORD="neo4jtest" -e HF_TOKEN=$HF_TOKEN -e MAX_INPUT_TOKENS=$MAX_INPUT_TOKENS -e LOGFLAG=True \
+        -e DATAPREP_COMPONENT_NAME="OPEA_DATAPREP_NEO4J_LLAMAINDEX" opea/dataprep-neo4j-llamaindex:comps
 
     sleep 1m
 
@@ -118,7 +117,7 @@ function validate_service() {
 function validate_microservice() {
     # validate neo4j-apoc
     validate_service \
-        "${host_ip}:${NEO4J_PORT1}" \
+        "${host_ip}:7474" \
         "200 OK" \
         "neo4j-apoc" \
         "neo4j-apoc" \
diff --git a/tests/retrievers/test_retrievers_opensearch.sh b/tests/retrievers/test_retrievers_opensearch.sh
index ae49c41a90..e8bb64850f 100644
--- a/tests/retrievers/test_retrievers_opensearch.sh
+++ b/tests/retrievers/test_retrievers_opensearch.sh
@@ -14,6 +14,7 @@ WORKPATH=$(dirname "$PWD")
 LOG_PATH="$WORKPATH/tests"
 export host_ip=$(hostname -I | awk '{print $1}')
 service_name="opensearch-vector-db tei-embedding-serving retriever-opensearch"
+retriever_service_name="retriever-opensearch"
 
 function build_docker_images() {
     cd $WORKPATH
@@ -34,6 +35,7 @@ function start_service() {
     export OPENSEARCH_INITIAL_ADMIN_PASSWORD="StRoNgOpEa0)"
     export EMBEDDING_MODEL_ID="BAAI/bge-base-en-v1.5"
     export TEI_EMBEDDING_ENDPOINT="http://${host_ip}:${TEI_EMBEDDER_PORT}"
+    export HUGGINGFACEHUB_API_TOKEN=${HF_TOKEN}
     export OPENSEARCH_URL="http://${host_ip}:${OPENSEARCH_PORT1}"
     export INDEX_NAME="file-index"
 
@@ -60,13 +62,13 @@ function validate_microservice() {
             echo "[ retriever ] Content is as expected."
         else
             echo "[ retriever ] Content does not match the expected result: $CONTENT"
-            docker logs ${service_name} >> ${LOG_PATH}/retriever.log
+            docker logs ${retriever_service_name} >> ${LOG_PATH}/retriever.log
             docker logs tei-embedding-serving >> ${LOG_PATH}/tei.log
             exit 1
         fi
     else
         echo "[ retriever ] HTTP status is not 200. Received status was $HTTP_STATUS"
-        docker logs ${service_name} >> ${LOG_PATH}/retriever.log
+        docker logs ${retriever_service_name} >> ${LOG_PATH}/retriever.log
         docker logs tei-embedding-serving >> ${LOG_PATH}/tei.log
         exit 1
     fi
@@ -80,14 +82,14 @@ function stop_docker() {
 }
 
 function main() {
-    stop_service
+    stop_docker
 
     build_docker_images
     start_service
 
     validate_microservice
 
-    stop_service
+    stop_docker
     echo y | docker system prune
 }
 
diff --git a/tests/retrievers/test_retrievers_vdms.sh b/tests/retrievers/test_retrievers_vdms.sh
index 732fd91134..42a1d550e5 100644
--- a/tests/retrievers/test_retrievers_vdms.sh
+++ b/tests/retrievers/test_retrievers_vdms.sh
@@ -37,7 +37,7 @@ function start_service() {
     export TEI_EMBEDDING_ENDPOINT="http://${host_ip}:${TEI_EMBEDDER_PORT}"
     export INDEX_NAME="rag-vdms"
     export VDMS_USE_CLIP=0 #set to 1 if openai clip embedding should be used
-    export HF_TOKEN=${HF_TOKEN}
+    export HUGGINGFACEHUB_API_TOKEN=${HF_TOKEN}
     export LOGFLAG=True
 
     cd $WORKPATH/comps/retrievers/deployment/docker_compose