diff --git a/README.md b/README.md
index 518dae4..30edbed 100644
--- a/README.md
+++ b/README.md
@@ -105,10 +105,10 @@ import litellm
 
 litellm.api_base = "http://localhost:8080"
 ```
-### Cache-aside
 
-Install with:
 
+### Cache-aside
+Install with:
 ```bash
 pip install semcache
 ```
@@ -127,6 +127,28 @@ response = client.get("Tell me France's capital city.")
 print(response)  # "Paris"
 ```
 
+
+or in Node.js
+
+Install with
+```bash
+npm install semcache
+```
+Use the sdk in your service
+
+```javascript
+const SemcacheClient = require('semcache');
+
+const client = new SemcacheClient('http://localhost:8080');
+
+(async () => {
+  await client.put('What is the capital of France?', 'Paris');
+
+  const result = await client.get('What is the capital of France?');
+  console.log(result); // => 'Paris'
+})();
+```
+
 ## Configuration
 
 Configure via environment variables or `config.yaml`:
diff --git a/docs/semcache/docs/getting-started.md b/docs/semcache/docs/getting-started.md
index 81a509f..5a9744b 100644
--- a/docs/semcache/docs/getting-started.md
+++ b/docs/semcache/docs/getting-started.md
@@ -19,7 +19,7 @@ docker run -p 8080:8080 semcache/semcache:latest
 
 Semcache will start on `http://localhost:8080` and is ready to proxy LLM requests.
 
-## Setting Up The Client
+## Setting up proxy client
 
 Semcache acts as a drop-in replacement for LLM APIs. Point your existing SDK to Semcache instead of the provider's endpoint:
 
@@ -105,7 +105,7 @@ This request will:
 3. The provider responds with the answer
 4. Semcache caches the response and returns it to you
 
-## Testing Semantic Similarity
+### Testing Semantic Similarity
 
 Now try a semantically similar but differently worded question:
 
@@ -153,7 +153,7 @@ Now try a semantically similar but differently worded question:
 
 Even though the wording is different, Semcache recognizes the semantic similarity and returns the cached response instantly - no API call to the upstream provider!
 
-## Checking Cache Status
+### Checking Cache Status
 
 You can verify cache hits by checking the response headers. If there is a cache hit the `X-Cache-Status` header will be set to `hit`:
 
@@ -249,6 +249,50 @@ You can verify cache hits by checking the response headers. If there is a cache
   </TabItem>
 </Tabs>
 
+
+## Setting up cache aside instance
+
+<Tabs groupId="sdk">
+  <TabItem value="python" label="Python" default>
+    Install with
+    ```bash
+    pip install semcache
+    ```
+
+    ```python
+    from semcache import Semcache
+    
+    # Initialize the client
+    client = Semcache(base_url="http://localhost:8080")
+    
+    # Store a key-data pair
+    client.put("What is the capital of France?", "Paris")
+    
+    # Retrieve data by semantic similarity
+    response = client.get("Tell me France's capital city.")
+    print(response)  # "Paris"
+    ```
+  </TabItem>
+  <TabItem value="Node.js" label="Node.js">
+    Install with
+    ```bash
+    npm install semcache
+    ```
+    ```javascript
+    const SemcacheClient = require('semcache');
+    
+    const client = new SemcacheClient('http://localhost:8080');
+    
+    (async () => {
+      await client.put('What is the capital of France?', 'Paris');
+    
+      const result = await client.get('What is the capital of France?');
+      console.log(result); // => 'Paris'
+    })();
+    ```
+  </TabItem>
+</Tabs>
+
 ## Monitor Your Cache
 
 Visit the built-in admin dashboard at `http://localhost:8080/admin` to monitor:
@@ -263,4 +307,4 @@ The process is identical across all providers - Semcache automatically detects t
 
 - **[LLM Providers & Tools](./llm-providers-tools.md)** - Configure additional providers like DeepSeek, Mistral, and custom LLMs
 - **[Configuration](./configuration/cache-settings.md)** - Adjust similarity thresholds and cache behavior  
-- **[Monitoring](./monitoring/metrics.md)** - Set up production monitoring with Prometheus and Grafana
\ No newline at end of file
+- **[Monitoring](./monitoring/metrics.md)** - Set up production monitoring with Prometheus and Grafana