Update README.md

README.md

@@ -30,6 +30,7 @@ This document explains how to call the **`generate_squirrel`** endpoint.
   |-----------|--------|----------|-----------------------------------------------------------------------------|
   | `text`    | string | Yes      | Text you want to synthesize. Limited to **300** characters.              |
   | `voice_id`| string | Yes      | **UUID** of the voice to use. Must match one of the IDs in the voices provided below.    |
+  | `return_raw_audio`| boolean | No **(default False)**      | If the endpoint should return raw audio bytes. Guide to use the bytes as wav output is shared below    |
 
 
 ## 3. Example Requests
@@ -52,6 +53,37 @@ This document explains how to call the **`generate_squirrel`** endpoint.
   response = requests.post(url, json=payload, headers=headers)
 
   ```
+
+  **Incase someone wants to use the raw audio for realtime usecase**
+  ```python
+    import time
+    import wave
+    import requests
+
+
+    # Audio parameters from the API headers
+    sample_rate = 24000
+    channels = 1
+    sample_width = 2  # 16-bit = 2 bytes
+
+    # Save raw PCM data as WAV
+    def save_pcm_as_wav(pcm_bytes: bytes, output_path: str):
+      with wave.open(output_path, 'wb') as wav_file:
+          wav_file.setnchannels(channels)
+          wav_file.setsampwidth(sample_width)
+          wav_file.setframerate(sample_rate)
+          wav_file.writeframes(pcm_bytes)
+
+    s = time.time()
+    response = requests.post("https://prod-api2.desivocal.com/dv/api/v0/tts_api/generate_squirrel", json={
+      "text": "Hello दोस्तों! Welcome to Ringg TTS. यह एक बहुत ही शानदार text to speech system है जो Hindi और English दोनों languages को support करता है।",
+      "voice_id": "83ba74e4-9efb-4db3-913a-f2a0ad66904d",
+      "return_raw_audio": True
+    })
+
+    save_pcm_as_wav(response.content, "output.wav")
+    print(time.time()-s)
+  ```
   
 ## 4. Voices Catalogue (CSV)