filler phrase while starting tool call (#232)

* filler phrase while starting tool call

* skip tts phrase when lang switches

* removed periods from filler phrases

* added env var ENABLE_FILLER_PHRASES_BEFORE_TOOL_CALL

* made step to 0.01 for voice provider config

Author: rootflo-hardik

wavefront/client/src/config/voice-providers.ts

@@ -70,23 +70,23 @@ export const VOICE_PROVIDERS_CONFIG: VoiceProvidersConfig = {
             default: 0.5,
             min: 0,
             max: 1,
-            step: 0.05,
+            step: 0.01,
             description: 'Controls voice consistency (0-1)',
           },
           similarity_boost: {
             type: 'number',
             default: 0.75,
             min: 0,
             max: 1,
-            step: 0.05,
+            step: 0.01,
             description: 'Enhances voice similarity (0-1)',
           },
           style: {
             type: 'number',
             default: 0.0,
             min: 0,
             max: 1,
-            step: 0.05,
+            step: 0.01,
             description: 'Style exaggeration (0-1)',
           },
           use_speaker_boost: {
@@ -99,7 +99,7 @@ export const VOICE_PROVIDERS_CONFIG: VoiceProvidersConfig = {
             default: 1.0,
             min: 0.25,
             max: 4.0,
-            step: 0.05,
+            step: 0.01,
             description: 'Speech speed (0.25-4.0)',
           },
         },
@@ -155,7 +155,7 @@ export const VOICE_PROVIDERS_CONFIG: VoiceProvidersConfig = {
             default: undefined,
             description: 'Speech speed multiplier',
             placeholder: '1.0',
-            step: 0.1,
+            step: 0.01,
           },
         },
       },
@@ -183,23 +183,23 @@ export const VOICE_PROVIDERS_CONFIG: VoiceProvidersConfig = {
             default: 0.0,
             min: -0.75,
             max: 0.75,
-            step: 0.05,
+            step: 0.01,
             description: 'Voice pitch (-0.75 to 0.75)',
           },
           pace: {
             type: 'number',
             default: 1.0,
             min: 0.3,
             max: 3.0,
-            step: 0.1,
+            step: 0.01,
             description: 'Speech pace (0.3-3.0)',
           },
           loudness: {
             type: 'number',
             default: 1.0,
             min: 0.1,
             max: 3.0,
-            step: 0.1,
+            step: 0.01,
             description: 'Volume (0.1-3.0)',
           },
           enable_preprocessing: {
@@ -212,7 +212,7 @@ export const VOICE_PROVIDERS_CONFIG: VoiceProvidersConfig = {
             default: 0.6,
             min: 0.01,
             max: 1.0,
-            step: 0.05,
+            step: 0.01,
             description: 'Randomness for bulbul v3 (0.01-1.0)',
           },
         },

wavefront/server/apps/call_processing/call_processing/constants/filler_phrases.py

@@ -0,0 +1,33 @@
+FILLER_PHRASES = {
+    'en': [
+        'Please wait a moment',
+        'Just a second',
+        'Give me a moment',
+        'One moment please',
+    ],
+    'hi': ['कृपया एक पल प्रतीक्षा करें', 'बस एक सेकंड', 'मुझे एक पल दें', 'एक पल रुकिए'],
+    'ta': [
+        'தயவுசெய்து ஒரு நிமிடம் காத்திருக்கவும்',
+        'ஒரு நொடி',
+        'எனக்கு ஒரு நிமிடம் கொடுங்கள்',
+        'ஒரு நிமிடம்',
+    ],
+    'te': [
+        'దయచేసి ఒక్క క్షణం వేచి ఉండండి',
+        'ఒక్క నిమిషం',
+        'నాకు ఒక్క క్షణం ఇవ్వండి',
+        'ఒక్క క్షణం దయచేసి',
+    ],
+    'ml': [
+        'ദയവായി ഒരു നിമിഷം കാത്തിരിക്കൂ',
+        'ഒരു സെക്കൻഡ്',
+        'എനിക്ക് ഒരു നിമിഷം തരൂ',
+        'ഒരു നിമിഷം ദയവായി',
+    ],
+    'kn': [
+        'ದಯವಿಟ್ಟು ಒಂದು ಕ್ಷಣ ಕಾಯಿರಿ',
+        'ಒಂದು ಕ್ಷಣ',
+        'ನನಗೆ ಒಂದು ಕ್ಷಣ ನೀಡಿ',
+        'ಒಂದು ಕ್ಷಣ ದಯವಿಟ್ಟು',
+    ],
+}

wavefront/server/apps/call_processing/call_processing/services/pipecat_service.py

@@ -6,6 +6,8 @@
 
 from typing import Dict, Any, List
 from copy import deepcopy
+import os
+import random
 from call_processing.log.logger import logger
 from call_processing.services.tool_wrapper_service import ToolWrapperFactory
 from call_processing.utils import get_current_ist_time_str
@@ -57,6 +59,7 @@
 from call_processing.constants.language_config import (
     LANGUAGE_INSTRUCTIONS,
 )
+from call_processing.constants.filler_phrases import FILLER_PHRASES
 
 
 class STTLanguageSwitcher(ParallelPipeline):
@@ -519,6 +522,25 @@ async def run_conversation(
         task_container['task'] = task
 
         # Register event handlers
+        @llm.event_handler('on_function_calls_started')
+        async def on_function_calls_started(service, function_calls):
+            if (
+                os.getenv('ENABLE_FILLER_PHRASES_BEFORE_TOOL_CALL', '').lower()
+                != 'true'
+            ):
+                return
+            # Skip filler phrase when language is switching — the TTS service's language
+            # may change before the queued frame is processed, causing a language mismatch error.
+            call_names = [fc.function_name for fc in function_calls]
+            if 'detect_and_switch_language' in call_names:
+                return
+            current_lang = language_state.get('current_language', 'en')
+            phrases = FILLER_PHRASES.get(current_lang)
+            if not phrases:
+                return
+            phrase = random.choice(phrases)
+            await task.queue_frame(TTSSpeakFrame(phrase))
+
         @transport.event_handler('on_client_connected')
         async def on_client_connected(transport, client):
             logger.info(f"Client connected for agent: {agent_config['name']}")