diff --git a/examples/deepseek/.env.example b/examples/deepseek/.env.example
new file mode 100644
index 00000000..a3994882
--- /dev/null
+++ b/examples/deepseek/.env.example
@@ -0,0 +1 @@
+DEEPSEEK_API_KEY=your_deepseek_api_key_here
diff --git a/examples/deepseek/README.md b/examples/deepseek/README.md
new file mode 100644
index 00000000..77a956fe
--- /dev/null
+++ b/examples/deepseek/README.md
@@ -0,0 +1,42 @@
+# Memori + DeepSeek Example
+
+This example demonstrates how to use Memori with DeepSeek AI.
+
+## Setup
+
+1. Install dependencies:
+```bash
+pip install -r requirements.txt
+```
+
+2. Set up your DeepSeek API key:
+```bash
+export DEEPSEEK_API_KEY="your-api-key"
+```
+
+Or create a `.env` file:
+```bash
+cp .env.example .env
+# Edit .env and add your API key
+```
+
+3. Run the example:
+```bash
+python main.py
+```
+
+## How It Works
+
+This example shows Memori's memory capabilities with DeepSeek:
+
+1. **First conversation**: Establishes facts (favorite color, city)
+2. **Second conversation**: DeepSeek automatically recalls the favorite color
+3. **Third conversation**: DeepSeek remembers the city
+
+Memori automatically captures and stores conversation context, making it available for future interactions.
+
+## Requirements
+
+- Python 3.10+
+- DeepSeek API key
+- Memori Python SDK
diff --git a/examples/deepseek/main.py b/examples/deepseek/main.py
new file mode 100644
index 00000000..76b77285
--- /dev/null
+++ b/examples/deepseek/main.py
@@ -0,0 +1,60 @@
+# Quickstart: Memori + DeepSeek + SQLite
+
+# Demonstrates how Memori adds memory across conversations with DeepSeek.
+
+import os
+from dotenv import load_dotenv
+
+load_dotenv()
+
+# Note: DeepSeek uses OpenAI-compatible API
+from openai import OpenAI
+from sqlalchemy import create_engine
+from sqlalchemy.orm import sessionmaker
+
+from memori import Memori
+
+# Setup DeepSeek client (OpenAI-compatible API)
+client = OpenAI(
+    api_key=os.getenv("DEEPSEEK_API_KEY", "<your_deepseek_api_key_here>"),
+    base_url="https://api.deepseek.com/v1",
+)
+
+# Setup SQLite
+engine = create_engine("sqlite:///deepseek_memori.db")
+Session = sessionmaker(bind=engine)
+
+# Setup Memori with DeepSeek
+mem = Memori(conn=Session).deepseek.register(client)
+mem.attribution(entity_id="user-123", process_id="deepseek-app")
+mem.config.storage.build()
+
+if __name__ == "__main__":
+    # First conversation - establish facts
+    print("背景信息：我叫berry peng，是一名agent开发工程师，目前住在上海")
+    response1 = client.chat.completions.create(
+        model="deepseek-chat",
+        messages=[
+            {
+                "role": "user",
+                "content": "我叫berry peng，是一名agent开发工程师，目前住在上海",
+            }
+        ],
+    )
+    print(f"AI: {response1.choices[0].message.content}\n")
+
+    # Second conversation - Memori recalls context automatically
+    print("You: 我住在哪？")
+    response2 = client.chat.completions.create(
+        model="deepseek-chat",
+        messages=[{"role": "user", "content": "我住在哪？"}],
+    )
+    print(f"AI: {response2.choices[0].message.content}\n")
+
+    # Third conversation - context is maintained
+    print("You: 我的职业是什么")
+    response3 = client.chat.completions.create(
+        model="deepseek-chat",
+        messages=[{"role": "user", "content": "我的职业是什么"}],
+    )
+    print(f"AI: {response3.choices[0].message.content}")
diff --git a/examples/deepseek/pyproject.toml b/examples/deepseek/pyproject.toml
new file mode 100644
index 00000000..363b1d0f
--- /dev/null
+++ b/examples/deepseek/pyproject.toml
@@ -0,0 +1,9 @@
+[project]
+name = "deepseek-example"
+version = "0.1.0"
+description = "Memori + DeepSeek example"
+dependencies = [
+    "memori",
+    "openai>=1.0.0",
+    "sqlalchemy>=2.0.0",
+]
diff --git a/examples/deepseek/requirements.txt b/examples/deepseek/requirements.txt
new file mode 100644
index 00000000..0c41cf0a
--- /dev/null
+++ b/examples/deepseek/requirements.txt
@@ -0,0 +1,3 @@
+memori
+openai>=1.0.0
+sqlalchemy>=2.0.0
diff --git a/memori/__init__.py b/memori/__init__.py
index 59d95e88..629182d6 100644
--- a/memori/__init__.py
+++ b/memori/__init__.py
@@ -21,6 +21,7 @@
 from memori.llm._providers import Anthropic as LlmProviderAnthropic
 from memori.llm._providers import Google as LlmProviderGoogle
 from memori.llm._providers import LangChain as LlmProviderLangChain
+from memori.llm._providers import DeepSeek as LlmProviderDeepSeek
 from memori.llm._providers import OpenAi as LlmProviderOpenAi
 from memori.llm._providers import PydanticAi as LlmProviderPydanticAi
 from memori.llm._providers import XAi as LlmProviderXAi
@@ -80,6 +81,7 @@ def __init__(self, conn: Callable[[], Any] | Any | None = None):
         self.llm = LlmRegistry(self)
         self.agno = LlmProviderAgno(self)
         self.anthropic = LlmProviderAnthropic(self)
+        self.deepseek = LlmProviderDeepSeek(self)
         self.google = LlmProviderGoogle(self)
         self.langchain = LlmProviderLangChain(self)
         self.openai = LlmProviderOpenAi(self)
diff --git a/memori/llm/__init__.py b/memori/llm/__init__.py
index 3758faaf..8f410c41 100644
--- a/memori/llm/__init__.py
+++ b/memori/llm/__init__.py
@@ -10,6 +10,6 @@
 
 from memori.llm import _clients  # noqa: F401
 from memori.llm._registry import Registry
-from memori.llm.adapters import anthropic, bedrock, google, openai, xai  # noqa: F401
+from memori.llm.adapters import anthropic, bedrock, deepseek, google, openai, xai  # noqa: F401
 
 __all__ = ["Registry"]
diff --git a/memori/llm/_base.py b/memori/llm/_base.py
index 5e5ea8e7..f7e76fd2 100644
--- a/memori/llm/_base.py
+++ b/memori/llm/_base.py
@@ -27,6 +27,7 @@
     agno_is_xai,
     llm_is_anthropic,
     llm_is_bedrock,
+    llm_is_deepseek,
     llm_is_google,
     llm_is_openai,
     llm_is_xai,
@@ -322,6 +323,7 @@ def inject_conversation_messages(self, kwargs: dict) -> dict:
             llm_is_openai(self.config.framework.provider, self.config.llm.provider)
             or agno_is_openai(self.config.framework.provider, self.config.llm.provider)
             or agno_is_xai(self.config.framework.provider, self.config.llm.provider)
+            or llm_is_deepseek(self.config.framework.provider, self.config.llm.provider)
         ):
             kwargs["messages"] = messages + kwargs["messages"]
         elif (
diff --git a/memori/llm/_clients.py b/memori/llm/_clients.py
index 3a99bf79..a35ba4b5 100644
--- a/memori/llm/_clients.py
+++ b/memori/llm/_clients.py
@@ -449,6 +449,63 @@ def register(self, client, _provider=None, stream=False):
         return self
 
 
+@Registry.register_client(lambda client: "deepseek" in str(type(client).__module__).lower())
+class DeepSeek(BaseClient):
+    def register(self, client, _provider=None, stream=False):
+        from memori.llm._constants import DEEPSEEK_LLM_PROVIDER
+
+        if not hasattr(client, "chat"):
+            raise RuntimeError("client provided is not instance of DeepSeek")
+
+        if not hasattr(client, "_memori_installed"):
+            client.chat._completions_create = client.chat.completions.create
+
+            try:
+                asyncio.get_running_loop()
+
+                if stream is True:
+                    client.chat.completions.create = (
+                        InvokeAsyncStream(
+                            self.config,
+                            client.chat._completions_create,
+                        )
+                        .set_client(_provider, DEEPSEEK_LLM_PROVIDER, getattr(client, "_version", None))
+                        .invoke
+                    )
+                else:
+                    client.chat.completions.create = (
+                        InvokeAsync(
+                            self.config,
+                            client.chat._completions_create,
+                        )
+                        .set_client(_provider, DEEPSEEK_LLM_PROVIDER, getattr(client, "_version", None))
+                        .invoke
+                    )
+            except RuntimeError:
+                if stream is True:
+                    client.chat.completions.create = (
+                        InvokeStream(
+                            self.config,
+                            client.chat._completions_create,
+                        )
+                        .set_client(_provider, DEEPSEEK_LLM_PROVIDER, getattr(client, "_version", None))
+                        .invoke
+                    )
+                else:
+                    client.chat.completions.create = (
+                        Invoke(
+                            self.config,
+                            client.chat._completions_create,
+                        )
+                        .set_client(_provider, DEEPSEEK_LLM_PROVIDER, getattr(client, "_version", None))
+                        .invoke
+                    )
+
+            client._memori_installed = True
+
+        return self
+
+
 @Registry.register_client(
     lambda client: type(client).__module__.startswith("pydantic_ai")
 )
diff --git a/memori/llm/_constants.py b/memori/llm/_constants.py
index 9811b5a4..d2eae841 100644
--- a/memori/llm/_constants.py
+++ b/memori/llm/_constants.py
@@ -14,6 +14,7 @@
 AGNO_OPENAI_LLM_PROVIDER = "openai"
 AGNO_XAI_LLM_PROVIDER = "xai"
 ATHROPIC_LLM_PROVIDER = "anthropic"
+DEEPSEEK_LLM_PROVIDER = "deepseek"
 GOOGLE_LLM_PROVIDER = "google"
 LANGCHAIN_CHATBEDROCK_LLM_PROVIDER = "chatbedrock"
 LANGCHAIN_CHATGOOGLEGENAI_LLM_PROVIDER = "chatgooglegenai"
diff --git a/memori/llm/_providers.py b/memori/llm/_providers.py
index 9fc6dafe..f06ac995 100644
--- a/memori/llm/_providers.py
+++ b/memori/llm/_providers.py
@@ -18,6 +18,7 @@
 from memori.llm._clients import OpenAi as OpenAiMemoriClient
 from memori.llm._clients import PydanticAi as PydanticAiMemoriClient
 from memori.llm._clients import XAi as XAiMemoriClient
+from memori.llm._clients import DeepSeek as DeepSeekMemoriClient
 
 
 class Agno(BaseProvider):
@@ -51,6 +52,14 @@ def register(self, client):
         return self.entity
 
 
+class DeepSeek(BaseProvider):
+    def register(self, client):
+        if self.client is None:
+            self.client = DeepSeekMemoriClient(self.config).register(client)
+
+        return self.entity
+
+
 class Google(BaseProvider):
     def register(self, client):
         warnings.warn(
diff --git a/memori/llm/_utils.py b/memori/llm/_utils.py
index 603a22e8..6e9f28ac 100644
--- a/memori/llm/_utils.py
+++ b/memori/llm/_utils.py
@@ -15,6 +15,7 @@
     AGNO_OPENAI_LLM_PROVIDER,
     AGNO_XAI_LLM_PROVIDER,
     ATHROPIC_LLM_PROVIDER,
+    DEEPSEEK_LLM_PROVIDER,
     GOOGLE_LLM_PROVIDER,
     LANGCHAIN_CHATBEDROCK_LLM_PROVIDER,
     LANGCHAIN_CHATGOOGLEGENAI_LLM_PROVIDER,
@@ -42,6 +43,10 @@ def llm_is_bedrock(provider, title):
     )
 
 
+def llm_is_deepseek(provider, title):
+    return title == DEEPSEEK_LLM_PROVIDER
+
+
 def llm_is_google(provider, title):
     return title == GOOGLE_LLM_PROVIDER or (
         provider_is_langchain(provider)
@@ -64,6 +69,10 @@ def agno_is_anthropic(provider, title):
     return provider_is_agno(provider) and title == AGNO_ANTHROPIC_LLM_PROVIDER
 
 
+def agno_is_deepseek(provider, title):
+    return provider_is_agno(provider) and title == DEEPSEEK_LLM_PROVIDER
+
+
 def agno_is_google(provider, title):
     return provider_is_agno(provider) and title == AGNO_GOOGLE_LLM_PROVIDER
 
diff --git a/memori/llm/adapters/deepseek/__init__.py b/memori/llm/adapters/deepseek/__init__.py
new file mode 100644
index 00000000..1b070d46
--- /dev/null
+++ b/memori/llm/adapters/deepseek/__init__.py
@@ -0,0 +1,13 @@
+r"""
+ __  __                           _
+|  \/  | ___ _ __ ___   ___  _ __(_)
+| |\/| |/ _ \ '_ ` _ \ / _ \| '__| |
+| |  | |  __/ | | | | | (_) | |  | |
+|_|  |_|\___|_| |_| |_|\___/|_|  |_|
+                  perfectam memoriam
+                       memorilabs.ai
+"""
+
+from ._adapter import Adapter
+
+__all__ = ["Adapter"]
diff --git a/memori/llm/adapters/deepseek/_adapter.py b/memori/llm/adapters/deepseek/_adapter.py
new file mode 100644
index 00000000..2dfc149f
--- /dev/null
+++ b/memori/llm/adapters/deepseek/_adapter.py
@@ -0,0 +1,93 @@
+r"""
+ __  __                           _
+|  \/  | ___ _ __ ___   ___  _ __(_)
+| |\/| |/ _ \ '_ ` _ \ / _ \| '__| |
+| |  | |  __/ | | | | | (_) | |  | |
+|_|  |_|\___|_| |_| |_|\___/|_|  |_|
+                  perfectam memoriam
+                       memorilabs.ai
+"""
+
+from memori.llm._base import BaseLlmAdaptor
+from memori.llm._registry import Registry
+from memori.llm._utils import llm_is_deepseek
+
+
+@Registry.register_adapter(llm_is_deepseek)
+class Adapter(BaseLlmAdaptor):
+    def get_formatted_query(self, payload):
+        """
+        [
+            {
+                "content": "...",
+                "role": "..."
+            }
+        ]
+        """
+
+        try:
+            messages = payload["conversation"]["query"].get("messages", [])
+            return self._exclude_injected_messages(messages, payload)
+        except KeyError:
+            return []
+
+    def get_formatted_response(self, payload):
+        try:
+            choices = payload["conversation"]["response"].get("choices", None)
+        except KeyError:
+            return []
+
+        response = []
+        if choices is not None:
+            if payload["conversation"]["query"].get("stream", None) is None:
+                # Unstreamed
+                # [
+                #   {
+                #       "finish_reason": "...",
+                #       "index": ...,
+                #       "message": {
+                #           "content": "...",
+                #           "role": "..."
+                #       }
+                #   }
+                # ]
+                for choice in choices:
+                    message = choice.get("message", None)
+                    if message is not None:
+                        content = message.get("content", None)
+                        if content is not None:
+                            response.append(
+                                {
+                                    "role": message["role"],
+                                    "text": content,
+                                    "type": "text",
+                                }
+                            )
+            else:
+                # Streamed
+                # [
+                #   {
+                #       "delta": {
+                #           "content": "...",
+                #           "role": "..."
+                #       }
+                #   }
+                # ]
+                content = []
+                role = None
+                for choice in choices:
+                    delta = choice.get("delta", None)
+                    if delta is not None:
+                        if role is None:
+                            role = delta.get("role", None)
+
+                        text_content = delta.get("content", None)
+                        if text_content is not None and len(text_content) > 0:
+                            content.append(text_content)
+
+                if len(content) > 0:
+                    response.append(
+                        {"role": role, "text": "".join(content), "type": "text"}
+                    )
+
+        return response