diff --git a/src/vocametrix/_generated/__init__.py b/src/vocametrix/_generated/__init__.py
index dafc615..883dd64 100644
--- a/src/vocametrix/_generated/__init__.py
+++ b/src/vocametrix/_generated/__init__.py
@@ -1,6 +1,6 @@
 # Auto-generated — do not edit. Run `python tasks.py regenerate` to update.
+"""A client library for accessing Vocametrix API"""
 
-""" A client library for accessing Vocametrix API """
 from .client import AuthenticatedClient, Client
 
 __all__ = (
diff --git a/src/vocametrix/_generated/models/post_analyze_phonemes_live_response_200.py b/src/vocametrix/_generated/models/post_analyze_phonemes_live_response_200.py
index db04a11..33756cb 100644
--- a/src/vocametrix/_generated/models/post_analyze_phonemes_live_response_200.py
+++ b/src/vocametrix/_generated/models/post_analyze_phonemes_live_response_200.py
@@ -31,8 +31,8 @@ class PostAnalyzePhonemesLiveResponse200:
         language (str | Unset): Echoed language code processed.
         transcription (str | Unset): Space-separated sequence of detected phonemes (empty string if silence).
         phoneme_count (float | Unset): Total number of phonemes detected.
-        phonemes (list[PostAnalyzePhonemesLiveResponse200PhonemesItem] | Unset): Array of the phoneme labels detected,
-            in order.
+        phonemes (list[PostAnalyzePhonemesLiveResponse200PhonemesItem] | Unset): Array of unique phoneme labels, in
+            order of first appearance (duplicates removed). Use `transcription` or `phoneme_timings` for the full sequence.
         phoneme_timings (list[PostAnalyzePhonemesLiveResponse200PhonemeTimingsItem] | Unset): Array of `{ phoneme,
             start, end, duration, confidence, second_phoneme, second_confidence }` — one entry per phoneme.
             `second_phoneme`/`second_confidence` (French champion models only) is the runner-up at that phoneme's most
diff --git a/src/vocametrix/_generated/models/post_language_chat_pronunciation_body.py b/src/vocametrix/_generated/models/post_language_chat_pronunciation_body.py
index 4825c17..06c0a90 100644
--- a/src/vocametrix/_generated/models/post_language_chat_pronunciation_body.py
+++ b/src/vocametrix/_generated/models/post_language_chat_pronunciation_body.py
@@ -23,7 +23,7 @@ class PostLanguageChatPronunciationBody:
     Attributes:
         message (str): REQUIRED. The learner's message in the target language.
         language (str): REQUIRED. Target language being learned.
-        age_level (str): REQUIRED. Categorical age/level label.
+        age_level (str): REQUIRED. CEFR level of the learner (A1, A2, B1, B2, C1, C2).
         topic (str): REQUIRED. Conversation topic.
         native_language (str | Unset): Optional. The learner's native language code.
         thread_id (str | Unset): Optional. Thread ID for multi-turn continuity.
diff --git a/src/vocametrix/_generated/models/post_language_chat_vocabulary_body.py b/src/vocametrix/_generated/models/post_language_chat_vocabulary_body.py
index d4b49cd..ad78d3d 100644
--- a/src/vocametrix/_generated/models/post_language_chat_vocabulary_body.py
+++ b/src/vocametrix/_generated/models/post_language_chat_vocabulary_body.py
@@ -19,7 +19,7 @@ class PostLanguageChatVocabularyBody:
         language (str): REQUIRED. Target language the user is learning (en-US, fr-FR, es-ES, de-DE, etc.)
         native_language (str): REQUIRED. The user's native language code — used to localize vocabulary hints and
             corrections
-        age_level (str): Age/level (child-beginner, teen-intermediate, adult-advanced, etc.)
+        age_level (str): CEFR level of the learner (A1, A2, B1, B2, C1, C2)
         topic (str): Conversation topic (family, travel, food, work, hobbies, etc.)
         thread_id (str | Unset): Optional thread ID to continue previous conversation
         email (str | Unset): Optional user email for tracking
diff --git a/tests/unit/test_client.py b/tests/unit/test_client.py
index afeefe1..8ef0a7c 100644
--- a/tests/unit/test_client.py
+++ b/tests/unit/test_client.py
@@ -197,8 +197,8 @@ def test_stuttering_raises_on_timeout(tmp_path, client):
 
 
 @respx.mock
-def test_client_email_used_as_namespace_default(tmp_path):
-    """email set at client level should appear in upload requests."""
+def test_client_email_not_sent_in_uploads(tmp_path):
+    """email is deprecated: a client-level email is no longer sent in upload requests."""
     client = VocametrixClient(api_key="key", email="user@example.com")
     wav = tmp_path / "vowel.wav"
     wav.write_bytes(b"RIFF" + b"\x00" * 40)
@@ -221,11 +221,11 @@ def capture_assign(request):
     client.avqi.calculate(sustained_vowel=str(wav))
     client.close()
 
-    assert captured_email.get("value") == "user@example.com"
+    assert "value" not in captured_email
 
 
 @respx.mock
-def test_per_call_email_overrides_client_default(tmp_path):
+def test_per_call_email_not_sent_in_uploads(tmp_path):
     client = VocametrixClient(api_key="key", email="client@example.com")
     wav = tmp_path / "vowel.wav"
     wav.write_bytes(b"RIFF" + b"\x00" * 40)
@@ -248,7 +248,7 @@ def capture_assign(request):
     client.dsi.calculate(sustained_vowel=str(wav), email="override@example.com")
     client.close()
 
-    assert captured_email.get("value") == "override@example.com"
+    assert "value" not in captured_email
 
 
 @pytest.mark.asyncio
@@ -315,7 +315,7 @@ def test_transcription_event_importable_from_top_level():
 
 
 @respx.mock
-def test_default_email_is_info_vocametrix(tmp_path):
+def test_default_client_sends_no_email(tmp_path):
     wav = tmp_path / "v.wav"
     wav.write_bytes(b"RIFF" + b"\x00" * 40)
     captured = {}
@@ -335,7 +335,7 @@ def capture(request):
     with VocametrixClient(api_key="key") as c:
         c.avqi.calculate(sustained_vowel=str(wav))
 
-    assert captured.get("email") == "info@vocametrix.com"
+    assert "email" not in captured
 
 
 
@@ -427,12 +427,6 @@ def test_client_has_ai_agents_namespace():
         assert hasattr(c.ai_agents, "interpret_metrics")
 
 
-def test_client_has_speech_coaching_namespace():
-    with VocametrixClient(api_key="key") as c:
-        assert hasattr(c, "speech_coaching")
-        assert hasattr(c.speech_coaching, "analyze")
-        assert hasattr(c.speech_coaching, "analyze_batch")
-        assert hasattr(c.speech_coaching, "get_batch_result")
 
 
 @respx.mock
@@ -447,15 +441,3 @@ def test_ai_agents_therapy_plan():
         )
     assert result["recommendation"] == "rest your voice"
 
-
-@respx.mock
-def test_speech_coaching_analyze():
-    respx.post(f"{BASE}/api/coaching-analysis").mock(
-        return_value=httpx.Response(200, json={"score": 87.5})
-    )
-    with VocametrixClient(api_key="key") as c:
-        result = c.speech_coaching.analyze(
-            reference_audio_url="https://cdn.example.com/ref.wav",
-            learner_audio_url="https://cdn.example.com/learner.wav",
-        )
-    assert result["score"] == 87.5