diff --git a/app/_layout.tsx b/app/_layout.tsx
index 3c2be2d..15c88bf 100644
--- a/app/_layout.tsx
+++ b/app/_layout.tsx
@@ -12,6 +12,7 @@ import { AuthProvider, useAuth } from '@/contexts/AuthContext';
 import { BeneficiaryProvider } from '@/contexts/BeneficiaryContext';
 import { VoiceTranscriptProvider } from '@/contexts/VoiceTranscriptContext';
 import { VoiceCallProvider } from '@/contexts/VoiceCallContext';
+import { VoiceProvider } from '@/contexts/VoiceContext';
 import { LoadingSpinner } from '@/components/ui/LoadingSpinner';
 import { FloatingCallBubble } from '@/components/FloatingCallBubble';
 
@@ -67,7 +68,9 @@ export default function RootLayout() {
           <BeneficiaryProvider>
             <VoiceTranscriptProvider>
               <VoiceCallProvider>
-                <RootLayoutNav />
+                <VoiceProvider>
+                  <RootLayoutNav />
+                </VoiceProvider>
               </VoiceCallProvider>
             </VoiceTranscriptProvider>
           </BeneficiaryProvider>
diff --git a/contexts/VoiceContext.tsx b/contexts/VoiceContext.tsx
new file mode 100644
index 0000000..28b6036
--- /dev/null
+++ b/contexts/VoiceContext.tsx
@@ -0,0 +1,382 @@
+/**
+ * Voice Context - Local STT/TTS integration with WellNuo API
+ *
+ * Provides voice session management:
+ * - STT (Speech-to-Text) via expo-speech-recognition
+ * - API calls to WellNuo ask_wellnuo_ai
+ * - TTS (Text-to-Speech) via expo-speech
+ *
+ * Flow: User speaks → STT → API → Response → TTS → Continue listening
+ */
+
+import React, {
+  createContext,
+  useContext,
+  useState,
+  useCallback,
+  useRef,
+  ReactNode,
+} from 'react';
+import * as Speech from 'expo-speech';
+import { api } from '@/services/api';
+
+// WellNuo API configuration (same as chat.tsx)
+const API_URL = 'https://eluxnetworks.net/function/well-api/api';
+const WELLNUO_USER = 'anandk';
+const WELLNUO_PASSWORD = 'anandk_8';
+
+// Single deployment mode - sends only deployment_id (no beneficiary_names_dict)
+const SINGLE_DEPLOYMENT_MODE = true;
+
+// Keywords for question normalization (same as chat.tsx)
+const STATUS_KEYWORDS = [
+  /\bhow\s+is\b/i,
+  /\bhow'?s\b/i,
+  /\bhow\s+are\b/i,
+  /\btell\s+me\s+about\b/i,
+  /\bwhat'?s\s+up\s+with\b/i,
+  /\bupdate\s+on\b/i,
+  /\bstatus\b/i,
+  /\bdoing\b/i,
+  /\bfeeling\b/i,
+  /\bcheck\s+on\b/i,
+  /\bis\s+\w+\s+okay\b/i,
+  /\bis\s+\w+\s+alright\b/i,
+  /\bis\s+\w+\s+fine\b/i,
+  /\bokay\?\b/i,
+  /\balright\?\b/i,
+];
+
+const SUBJECT_KEYWORDS = [
+  /\bdad\b/i,
+  /\bfather\b/i,
+  /\bferdinand\b/i,
+  /\bhim\b/i,
+  /\bhe\b/i,
+  /\bmy\s+dad\b/i,
+  /\bmy\s+father\b/i,
+  /\bthe\s+patient\b/i,
+  /\bloved\s+one\b/i,
+  /\bparent\b/i,
+  /\bgrandpa\b/i,
+  /\bgrandfather\b/i,
+];
+
+/**
+ * Normalize question for WellNuo API (same logic as chat.tsx)
+ */
+function normalizeQuestion(userMessage: string): string {
+  const msgLower = userMessage.toLowerCase().trim();
+
+  const isStatusQuery = STATUS_KEYWORDS.some((pattern) => pattern.test(msgLower));
+  const isAboutRecipient = SUBJECT_KEYWORDS.some((pattern) => pattern.test(msgLower));
+
+  if (isStatusQuery && isAboutRecipient) {
+    console.log(`[VoiceContext] Normalized '${userMessage}' -> 'how is dad doing'`);
+    return 'how is dad doing';
+  }
+
+  if (isStatusQuery && !isAboutRecipient) {
+    console.log(
+      `[VoiceContext] Normalized '${userMessage}' -> 'how is dad doing' (assumed recipient)`
+    );
+    return 'how is dad doing';
+  }
+
+  console.log(`[VoiceContext] No normalization applied to: '${userMessage}'`);
+  return userMessage;
+}
+
+export type VoiceStatus = 'idle' | 'listening' | 'processing' | 'speaking';
+
+interface VoiceContextValue {
+  // Current status of the voice session
+  status: VoiceStatus;
+  // Whether voice session is active (not idle)
+  isActive: boolean;
+  // Whether STT is currently listening
+  isListening: boolean;
+  // Whether TTS is currently speaking
+  isSpeaking: boolean;
+  // Whether processing API request
+  isProcessing: boolean;
+  // Current/last transcript from STT
+  transcript: string;
+  // Partial transcript (real-time preview)
+  partialTranscript: string;
+  // Last API response
+  lastResponse: string | null;
+  // Error message if any
+  error: string | null;
+
+  // Start voice session (begin listening)
+  startSession: () => void;
+  // Stop voice session
+  stopSession: () => void;
+
+  // Send transcript to API and get response with TTS
+  // Called automatically when STT detects speech end, or manually
+  sendTranscript: (text: string) => Promise<string | null>;
+
+  // Update transcript from external STT hook
+  setTranscript: (text: string) => void;
+  setPartialTranscript: (text: string) => void;
+
+  // Set status from external STT/TTS hooks
+  setStatus: (status: VoiceStatus) => void;
+  setIsListening: (listening: boolean) => void;
+  setIsSpeaking: (speaking: boolean) => void;
+
+  // Speak text using TTS
+  speak: (text: string) => Promise<void>;
+  // Stop TTS
+  stopSpeaking: () => void;
+}
+
+const VoiceContext = createContext<VoiceContextValue | undefined>(undefined);
+
+export function VoiceProvider({ children }: { children: ReactNode }) {
+  const [status, setStatus] = useState<VoiceStatus>('idle');
+  const [transcript, setTranscript] = useState('');
+  const [partialTranscript, setPartialTranscript] = useState('');
+  const [lastResponse, setLastResponse] = useState<string | null>(null);
+  const [error, setError] = useState<string | null>(null);
+  const [isListening, setIsListening] = useState(false);
+  const [isSpeaking, setIsSpeaking] = useState(false);
+
+  // API token cache
+  const apiTokenRef = useRef<string | null>(null);
+
+  // Deployment ID from settings
+  const deploymentIdRef = useRef<string | null>(null);
+
+  // Load deployment ID on mount
+  React.useEffect(() => {
+    const loadDeploymentId = async () => {
+      const savedId = await api.getDeploymentId();
+      deploymentIdRef.current = savedId;
+      console.log('[VoiceContext] Loaded deployment ID:', savedId);
+    };
+    loadDeploymentId();
+  }, []);
+
+  /**
+   * Get WellNuo API token (same as chat.tsx)
+   */
+  const getWellNuoToken = useCallback(async (): Promise<string> => {
+    if (apiTokenRef.current) {
+      return apiTokenRef.current;
+    }
+
+    const nonce = Math.floor(Math.random() * 1000000).toString();
+    const response = await fetch(API_URL, {
+      method: 'POST',
+      headers: { 'Content-Type': 'application/x-www-form-urlencoded' },
+      body: new URLSearchParams({
+        function: 'credentials',
+        clientId: 'MA_001',
+        user_name: WELLNUO_USER,
+        ps: WELLNUO_PASSWORD,
+        nonce: nonce,
+      }).toString(),
+    });
+
+    const data = await response.json();
+    if (data.status === '200 OK' && data.access_token) {
+      apiTokenRef.current = data.access_token;
+      console.log('[VoiceContext] WellNuo token obtained');
+      return data.access_token;
+    }
+    throw new Error('Failed to authenticate with WellNuo API');
+  }, []);
+
+  /**
+   * Send transcript to WellNuo API and speak the response
+   */
+  const sendTranscript = useCallback(
+    async (text: string): Promise<string | null> => {
+      const trimmedText = text.trim();
+      if (!trimmedText) {
+        console.log('[VoiceContext] Empty transcript, skipping API call');
+        return null;
+      }
+
+      console.log('[VoiceContext] Sending transcript to API:', trimmedText);
+      setStatus('processing');
+      setError(null);
+
+      try {
+        // Get API token
+        const token = await getWellNuoToken();
+
+        // Normalize question
+        const normalizedQuestion = normalizeQuestion(trimmedText);
+
+        // Get deployment ID
+        const deploymentId = deploymentIdRef.current || '21';
+
+        // Build request params
+        const requestParams: Record<string, string> = {
+          function: 'ask_wellnuo_ai',
+          clientId: 'MA_001',
+          user_name: WELLNUO_USER,
+          token: token,
+          question: normalizedQuestion,
+          deployment_id: deploymentId,
+        };
+
+        // Only add beneficiary_names_dict if NOT in single deployment mode
+        if (!SINGLE_DEPLOYMENT_MODE) {
+          // For full app, would include beneficiary names dict
+          // Currently single deployment mode only
+        }
+
+        const response = await fetch(API_URL, {
+          method: 'POST',
+          headers: { 'Content-Type': 'application/x-www-form-urlencoded' },
+          body: new URLSearchParams(requestParams).toString(),
+        });
+
+        const data = await response.json();
+
+        if (data.ok && data.response?.body) {
+          const responseText = data.response.body;
+          console.log('[VoiceContext] API response:', responseText.slice(0, 100) + '...');
+          setLastResponse(responseText);
+
+          // Speak the response
+          await speak(responseText);
+
+          return responseText;
+        } else {
+          // Token might be expired
+          if (data.status === '401 Unauthorized') {
+            apiTokenRef.current = null;
+            throw new Error('Session expired, please try again');
+          }
+          throw new Error(data.message || 'Could not get response');
+        }
+      } catch (err) {
+        const errorMsg = err instanceof Error ? err.message : 'Unknown error';
+        console.error('[VoiceContext] API error:', errorMsg);
+        setError(errorMsg);
+        setStatus('idle');
+        return null;
+      }
+    },
+    [getWellNuoToken]
+  );
+
+  /**
+   * Speak text using TTS
+   */
+  const speak = useCallback(async (text: string): Promise<void> => {
+    if (!text.trim()) return;
+
+    console.log('[VoiceContext] Speaking:', text.slice(0, 50) + '...');
+    setStatus('speaking');
+    setIsSpeaking(true);
+
+    return new Promise((resolve) => {
+      Speech.speak(text, {
+        language: 'en-US',
+        rate: 0.9,
+        pitch: 1.0,
+        onStart: () => {
+          console.log('[VoiceContext] TTS started');
+        },
+        onDone: () => {
+          console.log('[VoiceContext] TTS completed');
+          setIsSpeaking(false);
+          // Return to listening state after speaking (if session is active)
+          setStatus('listening');
+          resolve();
+        },
+        onError: (error) => {
+          console.error('[VoiceContext] TTS error:', error);
+          setIsSpeaking(false);
+          setStatus('listening');
+          resolve();
+        },
+        onStopped: () => {
+          console.log('[VoiceContext] TTS stopped');
+          setIsSpeaking(false);
+          setStatus('listening');
+          resolve();
+        },
+      });
+    });
+  }, []);
+
+  /**
+   * Stop TTS playback
+   */
+  const stopSpeaking = useCallback(() => {
+    Speech.stop();
+    setIsSpeaking(false);
+  }, []);
+
+  /**
+   * Start voice session
+   */
+  const startSession = useCallback(() => {
+    console.log('[VoiceContext] Starting voice session');
+    setStatus('listening');
+    setIsListening(true);
+    setError(null);
+    setTranscript('');
+    setPartialTranscript('');
+  }, []);
+
+  /**
+   * Stop voice session
+   */
+  const stopSession = useCallback(() => {
+    console.log('[VoiceContext] Stopping voice session');
+    Speech.stop();
+    setStatus('idle');
+    setIsListening(false);
+    setIsSpeaking(false);
+    setError(null);
+  }, []);
+
+  // Computed values
+  const isActive = status !== 'idle';
+  const isProcessing = status === 'processing';
+
+  return (
+    <VoiceContext.Provider
+      value={{
+        status,
+        isActive,
+        isListening,
+        isSpeaking,
+        isProcessing,
+        transcript,
+        partialTranscript,
+        lastResponse,
+        error,
+        startSession,
+        stopSession,
+        sendTranscript,
+        setTranscript,
+        setPartialTranscript,
+        setStatus,
+        setIsListening,
+        setIsSpeaking,
+        speak,
+        stopSpeaking,
+      }}
+    >
+      {children}
+    </VoiceContext.Provider>
+  );
+}
+
+export function useVoice() {
+  const context = useContext(VoiceContext);
+  if (!context) {
+    throw new Error('useVoice must be used within VoiceProvider');
+  }
+  return context;
+}