diff --git a/graphiti_core/embedder/__init__.py b/graphiti_core/embedder/__init__.py
index aea15619..417aa598 100644
--- a/graphiti_core/embedder/__init__.py
+++ b/graphiti_core/embedder/__init__.py
@@ -1,8 +1,11 @@
 from .client import EmbedderClient
 from .openai import OpenAIEmbedder, OpenAIEmbedderConfig
+from .ollama import OllamaEmbedder, OllamaEmbedderConfig
 
 __all__ = [
     'EmbedderClient',
     'OpenAIEmbedder',
     'OpenAIEmbedderConfig',
+    'OllamaEmbedder',
+    'OllamaEmbedderConfig',
 ]
diff --git a/mcp_server/.env.example b/mcp_server/.env.example
index 1e70ee56..795b5a48 100644
--- a/mcp_server/.env.example
+++ b/mcp_server/.env.example
@@ -14,6 +14,17 @@ MODEL_NAME=gpt-4.1-mini
 # Optional: Only needed for non-standard OpenAI endpoints
 # OPENAI_BASE_URL=https://api.openai.com/v1
 
+# Embedder Configuration
+# Provider is auto-detected based on configuration:
+# - Azure: if AZURE_OPENAI_EMBEDDING_ENDPOINT is set
+# - Ollama: if USE_OLLAMA_FOR_EMBEDDER is set to true
+# - OpenAI: default (no additional config needed)
+# USE_OLLAMA_FOR_EMBEDDER=true  # Set this to true to use Ollama
+# OLLAMA_EMBEDDER_API_KEY=ollama  # Ollama API key (optional, defaults to 'ollama')
+# OLLAMA_EMBEDDER_BASE_URL=http://localhost:11434  # Ollama base URL (when using Ollama)
+# OLLAMA_EMBEDDER_MODEL_NAME=nomic-embed-text  # Ollama embedding model to use
+# OLLAMA_EMBEDDER_DIMENSION=768  # Ollama embedding dimension (model-specific)
+
 # Optional: Group ID for namespacing graph data
 # GROUP_ID=my_project
 
diff --git a/mcp_server/.env.example.gemini_ollama b/mcp_server/.env.example.gemini_ollama
index 70c13935..9b68c355 100644
--- a/mcp_server/.env.example.gemini_ollama
+++ b/mcp_server/.env.example.gemini_ollama
@@ -18,10 +18,11 @@ OPENAI_BASE_URL=https://generativelanguage.googleapis.com/v1beta
 # Embedder Configuration
 # Optional: Separate API key and URL for embedder (falls back to OPENAI_API_KEY and OPENAI_BASE_URL if not set)
 # Note: OpenRouter does not support embeddings API, using Ollama as free alternative
-EMBEDDER_API_KEY=ollama
-EMBEDDER_BASE_URL=http://localhost:11434
-EMBEDDER_MODEL_NAME=nomic-embed-text
-EMBEDDER_DIMENSION=768
+USE_OLLAMA_FOR_EMBEDDER=true
+OLLAMA_EMBEDDER_API_KEY=ollama
+OLLAMA_EMBEDDER_BASE_URL=http://localhost:11434
+OLLAMA_EMBEDDER_MODEL_NAME=nomic-embed-text
+OLLAMA_EMBEDDER_DIMENSION=768
 
 # Optional: Group ID for namespacing graph data
 # GROUP_ID=my_project
diff --git a/mcp_server/.env.example.openrouter_ollama b/mcp_server/.env.example.openrouter_ollama
index 14181c3b..0939f2dc 100644
--- a/mcp_server/.env.example.openrouter_ollama
+++ b/mcp_server/.env.example.openrouter_ollama
@@ -17,10 +17,11 @@ OPENAI_BASE_URL=https://openrouter.ai/api/v1
 # Embedder Configuration
 # Optional: Separate API key and URL for embedder (falls back to OPENAI_API_KEY and OPENAI_BASE_URL if not set)
 # Note: OpenRouter does not support embeddings API, using Ollama as free alternative
-EMBEDDER_API_KEY=ollama
-EMBEDDER_BASE_URL=http://localhost:11434
-EMBEDDER_MODEL_NAME=nomic-embed-text
-EMBEDDER_DIMENSION=768
+USE_OLLAMA_FOR_EMBEDDER=true
+OLLAMA_EMBEDDER_API_KEY=ollama
+OLLAMA_EMBEDDER_BASE_URL=http://localhost:11434
+OLLAMA_EMBEDDER_MODEL_NAME=nomic-embed-text
+OLLAMA_EMBEDDER_DIMENSION=768
 
 # Optional: Group ID for namespacing graph data
 # GROUP_ID=my_project
diff --git a/mcp_server/README.md b/mcp_server/README.md
index d957feb8..ee52928b 100644
--- a/mcp_server/README.md
+++ b/mcp_server/README.md
@@ -92,6 +92,11 @@ The server uses the following environment variables:
 - `MODEL_NAME`: OpenAI model name to use for LLM operations.
 - `SMALL_MODEL_NAME`: OpenAI model name to use for smaller LLM operations.
 - `LLM_TEMPERATURE`: Temperature for LLM responses (0.0-2.0).
+- `USE_OLLAMA_FOR_EMBEDDER`: Set to `true` to use Ollama for embeddings (auto-detects Ollama provider)
+- `OLLAMA_EMBEDDER_API_KEY`: Ollama API key (optional, defaults to 'ollama')
+- `OLLAMA_EMBEDDER_BASE_URL`: Ollama base URL for embedder API (when using Ollama)
+- `OLLAMA_EMBEDDER_MODEL_NAME`: Ollama embedding model name
+- `OLLAMA_EMBEDDER_DIMENSION`: Ollama embedding dimension
 - `AZURE_OPENAI_ENDPOINT`: Optional Azure OpenAI LLM endpoint URL
 - `AZURE_OPENAI_DEPLOYMENT_NAME`: Optional Azure OpenAI LLM deployment name
 - `AZURE_OPENAI_API_VERSION`: Optional Azure OpenAI LLM API version
diff --git a/mcp_server/graphiti_mcp_server.py b/mcp_server/graphiti_mcp_server.py
index e8ef4da1..b0901b6d 100644
--- a/mcp_server/graphiti_mcp_server.py
+++ b/mcp_server/graphiti_mcp_server.py
@@ -355,11 +355,11 @@ class GraphitiEmbedderConfig(BaseModel):
 
     model: str = DEFAULT_EMBEDDER_MODEL
     api_key: str | None = None
-    provider: str = "openai"  # "openai", "ollama", or "azure"
     azure_openai_endpoint: str | None = None
     azure_openai_deployment_name: str | None = None
     azure_openai_api_version: str | None = None
     azure_openai_use_managed_identity: bool = False
+    use_ollama_for_embedder: bool = False
 
     @classmethod
     def from_env(cls) -> 'GraphitiEmbedderConfig':
@@ -370,14 +370,11 @@ class GraphitiEmbedderConfig(BaseModel):
         model = model_env if model_env.strip() else DEFAULT_EMBEDDER_MODEL
 
         # Get embedder-specific API key and base URL, fallback to general OpenAI settings
-        api_key = os.environ.get('EMBEDDER_API_KEY') or os.environ.get('OPENAI_API_KEY')
 
-        # Detect provider based on configuration
-        provider = "openai"  # default
-        if api_key and api_key.lower() == "ollama":
-            provider = "ollama"
-
-        logger.info(f'GraphitiEmbedderConfig provider: {provider}')
+        # Detect provider based on configuration (similar to Azure pattern)
+        use_ollama_for_embedder = (
+            os.environ.get('USE_OLLAMA_FOR_EMBEDDER', 'false').lower() == 'true'
+        )
 
         azure_openai_endpoint = os.environ.get('AZURE_OPENAI_EMBEDDING_ENDPOINT', None)
         azure_openai_api_version = os.environ.get('AZURE_OPENAI_EMBEDDING_API_VERSION', None)
@@ -415,12 +412,19 @@ class GraphitiEmbedderConfig(BaseModel):
                 api_key=api_key,
                 azure_openai_api_version=azure_openai_api_version,
                 azure_openai_deployment_name=azure_openai_deployment_name,
+                use_ollama_for_embedder=False,
             )
         else:
+            if use_ollama_for_embedder:
+                api_key_env = os.environ.get("OLLAMA_EMBEDDER_API_KEY")
+                api_key = api_key_env if api_key_env else 'ollama'
+                logger.info(f'ollama api_key: {api_key}')
+            else:
+                api_key = os.environ.get("OPENAI_API_KEY")
             return cls(
                 model=model,
-                api_key=os.environ.get('OPENAI_API_KEY'),
-                provider=provider,
+                api_key=api_key,
+                use_ollama_for_embedder=use_ollama_for_embedder,
             )
 
     def create_client(self) -> EmbedderClient | None:
@@ -452,16 +456,16 @@ class GraphitiEmbedderConfig(BaseModel):
             else:
                 logger.error('OPENAI_API_KEY must be set when using Azure OpenAI API')
                 return None
-        elif self.provider == "ollama":
+        elif self.use_ollama_for_embedder:
 
-            base_url_env = os.environ.get('EMBEDDER_BASE_URL')
+            base_url_env = os.environ.get('OLLAMA_EMBEDDER_BASE_URL')
             base_url = base_url_env if base_url_env else 'http://localhost:11434'
 
-            model_env = os.environ.get('EMBEDDER_MODEL_NAME')
+            model_env = os.environ.get('OLLAMA_EMBEDDER_MODEL_NAME')
             model = model_env if model_env else 'nomic-embed-text'
 
             # Get embedding dimension from environment
-            embedding_dim_env = os.environ.get('EMBEDDER_DIMENSION')
+            embedding_dim_env = os.environ.get('OLLAMA_EMBEDDER_DIMENSION')
             embedding_dim = int(embedding_dim_env) if embedding_dim_env else 768
 
             logger.info(f'ollama model: {model}')