feat: voice recording + TTS read-aloud for assistant messages

Audio recording: - Mic button wired to MediaRecorder (AAC/M4A, 44.1kHz/128kbps) - Tap to start, tap again to stop — auto-uploads as audio attachment - RECORD_AUDIO permission with runtime request - Recording state: red stop-square on mic button TTS read-aloud: - Volume button on assistant messages → POST /api/v1/audio-gen - Gemini-TTS generates speech server-side, MediaPlayer streams it - Tap again to stop playback, accent-colored icon when playing - Graceful error handling (Toast on failure) Supporting changes: - KaizenApi.generateSpeech() for /api/v1/audio-gen endpoint - KaizenIcons.Volume2 added (Lucide) - DE/EN strings for recording, permission, TTS states
2026-06-23 00:27:02 +02:00 · 2026-06-23 00:27:02 +02:00 · 9e1f30de1f
commit 9e1f30de1f
parent edd8ccaa05
7 changed files with 198 additions and 2 deletions
--- a/app/src/main/AndroidManifest.xml
+++ b/app/src/main/AndroidManifest.xml
@ -8,6 +8,9 @@
    <!-- Required to talk to the Kaizen backend (login + chat streaming) -->
    <uses-permission android:name="android.permission.INTERNET" />

+    <!-- Required for voice recording (mic button in chat) -->
+    <uses-permission android:name="android.permission.RECORD_AUDIO" />
+
    <application
        android:allowBackup="true"
        android:dataExtractionRules="@xml/data_extraction_rules"
--- a/app/src/main/java/dev/kaizen/app/chat/ChatComponents.kt
+++ b/app/src/main/java/dev/kaizen/app/chat/ChatComponents.kt
@ -331,6 +331,8 @@ fun MessageRow(
    onCopy: (String) -> Unit = {},
    onDelete: ((String) -> Unit)? = null,
    onRegenerate: ((String) -> Unit)? = null,
+    onReadAloud: ((String) -> Unit)? = null,
+    isPlayingTts: Boolean = false,
 ) {
    val cs = MaterialTheme.colorScheme
    val isDark = isSystemInDarkTheme()
@ -461,6 +463,14 @@ fun MessageRow(
                        copied = false
                    }
                }
+                onReadAloud?.let { readAloud ->
+                    MessageActionButton(
+                        icon = KaizenIcons.Volume2,
+                        tint = if (isPlayingTts) accent.primary else cs.onSurfaceVariant.copy(alpha = 0.35f),
+                    ) {
+                        haptics.tick(); readAloud(message.content)
+                    }
+                }
                onRegenerate?.let { regen ->
                    MessageActionButton(KaizenIcons.RefreshCw, cs.onSurfaceVariant.copy(alpha = 0.35f)) {
                        haptics.tick(); regen(message.wireId)
@ -527,6 +537,8 @@ fun ChatInput(
    enabled: Boolean,
    isStreaming: Boolean = false,
    onStop: () -> Unit = {},
+    isRecording: Boolean = false,
+    onMicClick: () -> Unit = {},
    pendingFiles: List<PendingFile> = emptyList(),
    onAddClick: () -> Unit = {},
    onRemoveFile: (String) -> Unit = {},
@ -584,11 +596,17 @@ fun ChatInput(
                }
                Spacer(Modifier.width(8.dp))
                Box(
-                    Modifier.size(38.dp).clip(KaizenShapes.circle).background(iconBg),
+                    Modifier.size(38.dp).clip(KaizenShapes.circle)
+                        .background(if (isRecording) Color(0xFFEF4444).copy(alpha = if (isDark) 0.85f else 0.80f) else iconBg)
+                        .clickable(onClick = onMicClick),
                    contentAlignment = Alignment.Center,
                ) {
+                    if (isRecording) {
+                        Icon(KaizenIcons.Square, stringResource(R.string.chat_recording), tint = Color.White, modifier = Modifier.size(14.dp))
+                    } else {
                        Icon(KaizenIcons.Mic, stringResource(R.string.chat_attachment), tint = cs.onSurfaceVariant, modifier = Modifier.size(19.dp))
                    }
+                }

                Spacer(Modifier.weight(1f))

--- a/app/src/main/java/dev/kaizen/app/chat/ChatScreen.kt
+++ b/app/src/main/java/dev/kaizen/app/chat/ChatScreen.kt
@ -82,10 +82,17 @@ import androidx.activity.result.contract.ActivityResultContracts
 import androidx.compose.ui.platform.LocalContext
 import androidx.compose.ui.res.stringResource
 import androidx.lifecycle.compose.LifecycleResumeEffect
+import android.Manifest
+import android.content.pm.PackageManager
+import android.media.MediaPlayer
+import android.media.MediaRecorder
+import android.widget.Toast
+import androidx.core.content.ContextCompat
 import kotlinx.coroutines.Job
 import kotlinx.coroutines.async
 import kotlinx.coroutines.launch
 import java.io.ByteArrayOutputStream
+import java.io.File

 // Screen enumeration for modular state-driven navigation (Zero Technical Debt)
 enum class AppScreen { Chat, Settings }
@ -111,6 +118,15 @@ fun ChatScreen(
    var chatModel by remember { mutableStateOf<String?>(null) }
    var streamJob by remember { mutableStateOf<Job?>(null) }

+    // Audio recording state
+    var isRecording by remember { mutableStateOf(false) }
+    val recorderRef = remember { mutableStateOf<MediaRecorder?>(null) }
+    val audioFileRef = remember { mutableStateOf<File?>(null) }
+
+    // TTS playback state
+    var playingTtsForId by remember { mutableStateOf<Long?>(null) }
+    val ttsPlayerRef = remember { mutableStateOf<MediaPlayer?>(null) }
+
    // Navigation State
    var currentScreen by remember { mutableStateOf(AppScreen.Chat) }

@ -191,6 +207,67 @@ fun ChatScreen(
        uploadPickedFile("photo_${System.currentTimeMillis()}.jpg", "image/jpeg", stream.toByteArray())
    }

+    val micPermission = rememberLauncherForActivityResult(ActivityResultContracts.RequestPermission()) { granted ->
+        if (granted) startRecording(context, recorderRef, audioFileRef) { isRecording = true }
+        else Toast.makeText(context, context.getString(R.string.chat_mic_permission), Toast.LENGTH_SHORT).show()
+    }
+
+    fun toggleRecording() {
+        if (isRecording) {
+            stopRecording(recorderRef, audioFileRef) { name, mime, bytes ->
+                isRecording = false
+                uploadPickedFile(name, mime, bytes)
+            }
+        } else {
+            if (ContextCompat.checkSelfPermission(context, Manifest.permission.RECORD_AUDIO) == PackageManager.PERMISSION_GRANTED) {
+                startRecording(context, recorderRef, audioFileRef) { isRecording = true }
+            } else {
+                micPermission.launch(Manifest.permission.RECORD_AUDIO)
+            }
+        }
+    }
+
+    fun readAloud(content: String, messageId: Long) {
+        val cfg = session.config ?: return
+        if (playingTtsForId == messageId) {
+            ttsPlayerRef.value?.release()
+            ttsPlayerRef.value = null
+            playingTtsForId = null
+            return
+        }
+        ttsPlayerRef.value?.release()
+        playingTtsForId = messageId
+        scope.launch {
+            val url = KaizenApi.generateSpeech(cfg.baseUrl, cfg.token, content)
+            if (url == null) {
+                playingTtsForId = null
+                Toast.makeText(context, context.getString(R.string.chat_tts_error), Toast.LENGTH_SHORT).show()
+                return@launch
+            }
+            try {
+                val player = MediaPlayer().apply {
+                    setDataSource(url)
+                    prepareAsync()
+                    setOnPreparedListener { start() }
+                    setOnCompletionListener {
+                        release()
+                        ttsPlayerRef.value = null
+                        playingTtsForId = null
+                    }
+                    setOnErrorListener { _, _, _ ->
+                        release()
+                        ttsPlayerRef.value = null
+                        playingTtsForId = null
+                        true
+                    }
+                }
+                ttsPlayerRef.value = player
+            } catch (_: Exception) {
+                playingTtsForId = null
+            }
+        }
+    }
+
    LaunchedEffect(session.config?.baseUrl, session.config?.token) {
        val cfg = session.config ?: return@LaunchedEffect
        loadError = null
@ -659,6 +736,10 @@ fun ChatScreen(
                                                send(userContent)
                                            }
                                        } else null,
+                                        onReadAloud = if (message.role == Role.Assistant && !message.streaming && message.content.isNotBlank()) { content ->
+                                            readAloud(content, message.id)
+                                        } else null,
+                                        isPlayingTts = playingTtsForId == message.id,
                                    )
                                }
                            }
@ -858,6 +939,8 @@ fun ChatScreen(
                                enabled = !isStreaming,
                                isStreaming = isStreaming,
                                onStop = { streamJob?.cancel() },
+                                isRecording = isRecording,
+                                onMicClick = { toggleRecording() },
                                pendingFiles = pendingFiles,
                                onAddClick = { showAttachMenu = !showAttachMenu },
                                onRemoveFile = { id -> pendingFiles.removeAll { it.id == id } },
@ -950,3 +1033,53 @@ private fun chatErrorText(e: Throwable, ctx: android.content.Context): String =
    e is ChatHttpException -> ctx.getString(R.string.error_server, e.code)
    else -> ctx.getString(R.string.error_no_connection)
 }
+
+@Suppress("DEPRECATION")
+private fun startRecording(
+    context: android.content.Context,
+    recorderRef: androidx.compose.runtime.MutableState<MediaRecorder?>,
+    audioFileRef: androidx.compose.runtime.MutableState<File?>,
+    onStarted: () -> Unit,
+) {
+    try {
+        val file = File(context.cacheDir, "voice_${System.currentTimeMillis()}.m4a")
+        val recorder = MediaRecorder(context).apply {
+            setAudioSource(MediaRecorder.AudioSource.MIC)
+            setOutputFormat(MediaRecorder.OutputFormat.MPEG_4)
+            setAudioEncoder(MediaRecorder.AudioEncoder.AAC)
+            setAudioSamplingRate(44100)
+            setAudioEncodingBitRate(128000)
+            setOutputFile(file.absolutePath)
+            prepare()
+            start()
+        }
+        recorderRef.value = recorder
+        audioFileRef.value = file
+        onStarted()
+    } catch (_: Exception) {
+        recorderRef.value = null
+        audioFileRef.value = null
+    }
+}
+
+private fun stopRecording(
+    recorderRef: androidx.compose.runtime.MutableState<MediaRecorder?>,
+    audioFileRef: androidx.compose.runtime.MutableState<File?>,
+    onRecorded: (String, String, ByteArray) -> Unit,
+) {
+    val recorder = recorderRef.value ?: return
+    val file = audioFileRef.value ?: return
+    try {
+        recorder.stop()
+        recorder.release()
+    } catch (_: Exception) {
+        recorder.release()
+    }
+    recorderRef.value = null
+    audioFileRef.value = null
+    if (file.exists() && file.length() > 0) {
+        val bytes = file.readBytes()
+        file.delete()
+        onRecorded("voice_${System.currentTimeMillis()}.m4a", "audio/mp4", bytes)
+    }
+}
--- a/app/src/main/java/dev/kaizen/app/net/KaizenApi.kt
+++ b/app/src/main/java/dev/kaizen/app/net/KaizenApi.kt
@ -572,6 +572,35 @@ object KaizenApi {
        }
    }

+    @Serializable private data class SpeechRequest(val prompt: String, val kind: String = "speech", val voice: String? = null)
+    @Serializable private data class SpeechResponse(val url: String = "", val cost: Double? = null)
+
+    /** POST /api/v1/audio-gen — generate speech from text via Gemini-TTS. Returns the audio URL or null. */
+    suspend fun generateSpeech(baseUrl: String, token: String, text: String, voice: String? = null): String? =
+        withContext(Dispatchers.IO) {
+            val body = json.encodeToString(SpeechRequest(
+                prompt = text.take(2000),
+                voice = voice,
+            )).toRequestBody(jsonMedia)
+            val req = Request.Builder()
+                .url("$baseUrl/api/v1/audio-gen")
+                .post(body)
+                .header("Authorization", "Bearer $token")
+                .build()
+            try {
+                client.newCall(req).execute().use { resp ->
+                    if (!resp.isSuccessful) {
+                        Log.w(TAG, "generateSpeech: HTTP ${resp.code}")
+                        return@use null
+                    }
+                    json.decodeFromString<SpeechResponse>(resp.body!!.string()).url.takeIf { it.isNotBlank() }
+                }
+            } catch (e: Exception) {
+                Log.w(TAG, "generateSpeech: ${e.javaClass.simpleName}: ${e.message}")
+                null
+            }
+        }
+
    suspend fun prewarm(baseUrl: String) = withContext(Dispatchers.IO) {
        try {
            val req = Request.Builder().url("$baseUrl/api/v1/meta").build()
--- a/app/src/main/java/dev/kaizen/app/ui/icon/KaizenIcons.kt
+++ b/app/src/main/java/dev/kaizen/app/ui/icon/KaizenIcons.kt
@ -91,6 +91,9 @@ object KaizenIcons {
    val ArrowDown: ImageVector get()     = Lucide.ArrowDown
    val ArrowUp: ImageVector get()       = Lucide.ArrowUp

+    // ── Audio ──────────────────────────────────────────────────────────────
+    val Volume2: ImageVector get()       = Lucide.Volume2
+
    // ── Custom (brand overrides) ────────────────────────────────────────────
    // Add hand-drawn SVGs here as ImageVector literals when ready.
    // Example:
--- a/app/src/main/res/values-en/strings.xml
+++ b/app/src/main/res/values-en/strings.xml
@ -19,6 +19,11 @@
    <string name="chat_error">Error</string>
    <string name="chat_open_menu">Open menu</string>
    <string name="chat_stop">Stop response</string>
+    <string name="chat_recording">Recording …</string>
+    <string name="chat_mic_permission">Microphone access required</string>
+    <string name="chat_read_aloud">Read aloud</string>
+    <string name="chat_tts_error">Read aloud failed</string>
+    <string name="chat_tts_unavailable">Read aloud not available (Vertex required)</string>
    <string name="attach_camera">Take Photo</string>
    <string name="attach_gallery">Photo &amp; Video</string>
    <string name="attach_file">Choose File</string>
--- a/app/src/main/res/values/strings.xml
+++ b/app/src/main/res/values/strings.xml
@ -21,6 +21,11 @@
    <string name="chat_error">Fehler</string>
    <string name="chat_open_menu">Menü öffnen</string>
    <string name="chat_stop">Antwort stoppen</string>
+    <string name="chat_recording">Aufnahme läuft …</string>
+    <string name="chat_mic_permission">Mikrofonzugriff wird benötigt</string>
+    <string name="chat_read_aloud">Vorlesen</string>
+    <string name="chat_tts_error">Vorlesen fehlgeschlagen</string>
+    <string name="chat_tts_unavailable">Vorlesen nicht verfügbar (Vertex benötigt)</string>
    <string name="attach_camera">Foto aufnehmen</string>
    <string name="attach_gallery">Foto &amp; Video</string>
    <string name="attach_file">Datei auswählen</string>