diff --git a/neuralspace/constants.py b/neuralspace/constants.py
index d3dff1f..fc883cb 100644
--- a/neuralspace/constants.py
+++ b/neuralspace/constants.py
@@ -38,6 +38,9 @@
 if env.VOCAB_ADAPT_URL is not None:
     VOCAB_ADAPT_URL = env.VOCAB_ADAPT_URL
 
+AMA_URL = 'api/v2/prompts'
+if env.AMA_URL is not None:
+    AMA_URL = env.AMA_URL
 
 # full url formation
 
@@ -46,6 +49,7 @@
 FULL_LANGS_URL = f'{BASE_URL.rstrip("/")}/{LANGS_URL.strip("/")}'
 FULL_VOICES_URL = f'{BASE_URL.rstrip("/")}/{VOICES_URL.strip("/")}'
 FULL_TOKEN_URL = f'{BASE_URL.rstrip("/")}/{TOKEN_URL.strip("/")}'
+FULL_AMA_URL = f'{BASE_URL.rstrip("/")}/{AMA_URL.strip("/")}'
 FULL_TTS_URL = f'{BASE_URL.rstrip("/")}/{TTS_URL.strip("/")}'
 FULL_VOCAB_ADAPT_URL = f'{BASE_URL.rstrip("/")}/{VOCAB_ADAPT_URL.strip("/")}'
 
diff --git a/neuralspace/env.py b/neuralspace/env.py
index e55777d..e5a5c06 100644
--- a/neuralspace/env.py
+++ b/neuralspace/env.py
@@ -10,6 +10,7 @@
 STREAM_URL = get('NS_STREAM_URL')
 LANGS_URL = get('NS_LANGS_URL')
 TOKEN_URL = get('NS_TOKEN_URL')
+AMA_URL = get('NS_AMA_URL')
 TIMEOUT_SEC = get('NS_TIMEOUT_SEC')
 TTS_URL = get('NS_TTS_URL')
 VOICES_URL = get('NS_VOICES_URL')
diff --git a/neuralspace/voice_ai.py b/neuralspace/voice_ai.py
index 1722fa0..1995799 100644
--- a/neuralspace/voice_ai.py
+++ b/neuralspace/voice_ai.py
@@ -270,6 +270,36 @@ def poll_until_complete(self, job_id: str, poll_schedule: Optional[List[float]]
         return result
 
 
+    def ama(self, job_id: str, prompts: List[str]) -> Dict[str, Any]:
+        """
+        Create and send a request for an AMA job.
+
+        Parameters
+        ----------
+        job_id: str
+            The job ID for the transcript on which to run AMA.
+        prompts: List[str]
+            List of prompts for the AMA.
+
+        Returns
+        -------
+        result: dict
+            The response from the server.
+        """
+        url = f'{K.FULL_AMA_URL.rstrip("/")}'
+        hdrs = self._create_headers()
+        hdrs['Content-Type'] = 'application/json'
+        data = json.dumps({
+            "jobId": job_id,
+            "prompts": prompts
+        })
+
+        sess = self._get_session()
+        r = sess.post(url, headers=hdrs, data=data)
+        resp = utils.get_json_resp(r)
+        return resp
+
+
     def _get_short_lived_token(self, timeout):
         url = f'{K.FULL_TOKEN_URL}?duration={timeout}'
         hdrs = self._create_headers()