refac

backend/open_webui/routers/audio/realtime.py

@@ -76,9 +76,15 @@ class CallProtocol:
 
     def observe(self, event):
         kind = event.get('type')
-        if kind == 'conversation.item.input_audio_transcription.completed':
-            if event.get('transcript', '').strip():
+        if kind in {
+            'conversation.item.input_audio_transcription.completed',
+            'conversation.item.input_audio_transcription.failed',
+        }:
+            if kind.endswith('.completed') and event.get('transcript', '').strip():
                 self.transcripts.add(event['item_id'])
+            else:
+                # Rejected turns must leave the provider's history as well as the chat.
+                return {'type': 'conversation.item.delete', 'item_id': event['item_id']}
         elif kind == 'response.created':
             self.responses.add(event['response']['id'])
             self.response_metadata[event['response']['id']] = event['response'].get('metadata') or {}
@@ -452,12 +458,14 @@ async def realtime_call(ws: WebSocket):
                         'audio': {
                             'input': {
                                 'format': {'type': 'audio/pcm', 'rate': 24000},
+                                'noise_reduction': {'type': 'near_field'},
                                 'transcription': {'model': config['audio.realtime.transcription_model']},
                                 # Wait for the finalized transcript before requesting a response. This gives
                                 # every function call an unambiguous input_item_id, even during barge-in.
                                 'turn_detection': {
                                     'type': 'server_vad',
-                                    'interrupt_response': True,
+                                    # The client pauses playback until speech is confirmed.
+                                    'interrupt_response': False,
                                     'create_response': False,
                                 },
                             },
@@ -502,7 +510,11 @@ async def realtime_call(ws: WebSocket):
                 if kind == 'error':
                     # Provider error text can include prompts or credentials.
                     raise ValueError('Voice provider rejected a request')
-                protocol.observe(event)
+                command = protocol.observe(event)
+                if command:
+                    # Clean up before the client can request another response.
+                    async with asyncio.timeout(5):
+                        await upstream.send_json(command)
                 if kind.startswith(('response.', 'conversation.item.', 'input_audio_buffer.')):
                     async with asyncio.timeout(5):
                         await ws.send_json(event)

backend/open_webui/utils/automations.py

@@ -505,6 +505,17 @@ async def execute_automation(app, automation: AutomationModel) -> None:
             'automation_id': automation.id,
             'background_tasks': {},
         }
+        # Match the chat UI's memory preference, including inherited defaults.
+        ui_settings = (user.settings.ui or {}) if user.settings else {}
+        memory = ui_settings.get('memory')
+        if memory is None:
+            defaults = await Config.get('ui.default_interface_settings') or {}
+            memory = defaults.get('memory')
+        if memory is None:
+            memory = await Config.get('memories.enable')
+        if memory:
+            form_data.setdefault('features', {})['memory'] = True
+
         # Call the full chat completion pipeline (same as POST /api/chat/completions).
         # The handler reference is stored on app.state to avoid circular imports.
         request = _build_request(app, token=token)

src/lib/components/channel/Messages/Message/UserStatus.svelte

@@ -1,5 +1,9 @@
 <script lang="ts">
 	import { getContext, onMount } from 'svelte';
+	import dayjs, { getDayjsLocale } from '$lib/dayjs';
+	import relativeTime from 'dayjs/plugin/relativeTime';
+
+	dayjs.extend(relativeTime);
 
 	const i18n = getContext<typeof import('$lib/i18n').default>('i18n');
 
@@ -30,9 +34,14 @@
 	let localTime = '';
 
 	$: lastSeen = user?.last_active_at
+		? dayjs(user.last_active_at * 1000)
+				.locale(getDayjsLocale($i18n.language))
+				.from(now)
+		: '';
+	$: lastSeenTimestamp = user?.last_active_at
 		? new Date(user.last_active_at * 1000).toLocaleString($i18n.language, {
 				dateStyle: 'medium',
-				timeStyle: 'short'
+				timeStyle: 'long'
 			})
 		: '';
 
@@ -166,7 +175,11 @@
 										<span class="sr-only">{$i18n.t('Last seen')}</span>
 									</Tooltip>
 								</dt>
-								<dd>{lastSeen}</dd>
+								<dd>
+									<Tooltip content={lastSeenTimestamp}>
+										<span>{lastSeen}</span>
+									</Tooltip>
+								</dd>
 							</div>
 						{/if}
 					</dl>

src/lib/utils/realtime-audio.js

@@ -10,6 +10,7 @@ class RealtimeAudioProcessor extends AudioWorkletProcessor {
 		this.queued = 0;
 		this.received = 0;
 		this.playing = false;
+		this.paused = false;
 		this.rendered = new Map();
 		this.ended = new Set();
 		this.ticks = 0;
@@ -22,6 +23,8 @@ class RealtimeAudioProcessor extends AudioWorkletProcessor {
 			if (data.type === 'capture') {
 				this.enabled = data.enabled;
 				this.captureLength = 0;
+			} else if (data.type === 'pause') {
+				this.paused = data.paused;
 			} else if (data.type === 'audio') {
 				// Speech can arrive faster than playback; release each chunk after it plays.
 				this.received += data.samples.length;
@@ -68,7 +71,7 @@ class RealtimeAudioProcessor extends AudioWorkletProcessor {
 		if (!this.playing && (this.queued >= 1920 || this.ended.has(this.queue[0]?.response_id))) {
 			this.playing = true;
 		}
-		if (this.playing) {
+		if (this.playing && !this.paused) {
 			let offset = 0;
 			while (offset < output.length && this.queue.length) {
 				const chunk = this.queue[0];

src/lib/utils/realtime.ts

@@ -350,9 +350,11 @@ export class RealtimeCall {
 			this.syncContext();
 			this.audio?.port.postMessage({ type: 'capture', enabled: !this.muted });
 		} else if (type === 'input_audio_buffer.speech_started') {
+			if (this.muted) return;
 			this.receivingSpeech = event.item_id;
 			this.userSpeaking = !this.muted;
-			this.stopSpeaking();
+			// VAD is provisional: retain the audio so a noise-only turn can resume it.
+			this.audio?.port.postMessage({ type: 'pause', paused: true });
 		} else if (type === 'input_audio_buffer.speech_stopped') {
 			if (this.receivingSpeech === event.item_id) this.userSpeaking = false;
 		} else if (
@@ -367,6 +369,8 @@ export class RealtimeCall {
 			if (this.inputs.has(event.item_id)) return;
 			const failed = type === 'conversation.item.input_audio_transcription.failed';
 			const text = failed ? '' : event.transcript?.trim();
+			if (text) this.stopSpeaking();
+			if (!this.receivingSpeech) this.audio?.port.postMessage({ type: 'pause', paused: false });
 			if (!text) {
 				// Empty VAD segments are not requests and must not produce assistant replies.
 				if (failed)
@@ -453,8 +457,7 @@ export class RealtimeCall {
 				if (this.animationCalls.has(event.item.call_id)) return;
 				this.animationCalls.add(event.item.call_id);
 				let status = 'unavailable';
-				if (!response || this.interrupted.has(event.response_id) || this.receivingSpeech)
-					status = 'cancelled';
+				if (!response || this.interrupted.has(event.response_id)) status = 'cancelled';
 				else {
 					try {
 						const args = JSON.parse(event.item.arguments);
@@ -713,12 +716,17 @@ export class RealtimeCall {
 		if (this.muted) {
 			this.inputLevel = 0;
 			this.userSpeaking = false;
+			this.receivingSpeech = '';
+			this.audio?.port.postMessage({ type: 'pause', paused: false });
 		}
 		this.stream?.getAudioTracks().forEach((track) => {
 			track.enabled = !this.muted;
 		});
 		this.audio?.port.postMessage({ type: 'capture', enabled: this.connected && !this.muted });
-		if (this.muted) this.send({ type: 'input_audio_buffer.clear' });
+		if (this.muted) {
+			this.send({ type: 'input_audio_buffer.clear' });
+			this.flush();
+		}
 		this.options.change();
 	}