fix: fix interruption handling for Gemini Live

1. Fixes #236
2. Fix run_inference for variable extraction for Gemini Live

api/services/pipecat/in_memory_buffers.py

@@ -120,9 +120,21 @@ class InMemoryLogsBuffer:
             f"Incremented turn counter to {self._turn_counter} for workflow {self._workflow_run_id}"
         )
 
+    @staticmethod
+    def _event_sort_key(event: dict) -> str:
+        payload_ts = event.get("payload", {}).get("timestamp")
+        return payload_ts or event.get("timestamp", "")
+
+    def _sorted_events(self) -> List[dict]:
+        # Stable sort by the realtime (payload) timestamp when available, falling
+        # back to the buffer-append timestamp. Python's sort is stable, so events
+        # sharing a key retain their original insertion order — this keeps
+        # consecutive bot-text chunks of a single turn contiguous.
+        return sorted(self._events, key=self._event_sort_key)
+
     def get_events(self) -> List[dict]:
-        """Get all events for final storage."""
-        return self._events
+        """Get all events for final storage, ordered by realtime timestamp."""
+        return self._sorted_events()
 
     def contains_user_speech(self) -> bool:
         """Return True if any final user transcription event has non-empty text."""
@@ -141,7 +153,7 @@ class InMemoryLogsBuffer:
         Filters for rtf-user-transcription (final) and rtf-bot-text events,
         formats them as '[timestamp] user/assistant: text\\n'.
         """
-        return _generate_transcript_text(self._events)
+        return _generate_transcript_text(self._sorted_events())
 
     def write_transcript_to_temp_file(self) -> Optional[str]:
         """Write transcript to a temporary text file and return the path.

api/services/pipecat/run_pipeline.py

@@ -616,10 +616,15 @@ async def _run_pipeline(
         llm = create_realtime_llm_service(user_config, audio_config)
         stt = None
         tts = None
+        # Realtime services don't implement run_inference, so create a
+        # separate text LLM for variable extraction and other out-of-band
+        # inference calls.
+        inference_llm = create_llm_service(user_config)
     else:
         stt = create_stt_service(user_config, audio_config, keyterms=keyterms)
         tts = create_tts_service(user_config, audio_config)
         llm = create_llm_service(user_config)
+        inference_llm = None
 
     workflow_graph = WorkflowGraph(ReactFlowDTO.model_validate(run_workflow_json))
 
@@ -703,9 +708,15 @@ async def _run_pipeline(
     context_compaction_enabled = (workflow.workflow_configurations or {}).get(
         "context_compaction_enabled", False
     )
+    # Context compaction doesn't apply in realtime mode: the speech-to-speech
+    # service manages its own conversation state server-side.
+    if is_realtime and context_compaction_enabled:
+        logger.info("Disabling context_compaction_enabled for realtime workflow run")
+        context_compaction_enabled = False
 
     engine = PipecatEngine(
         llm=llm,
+        inference_llm=inference_llm,
         workflow=workflow_graph,
         call_context_vars=merged_call_context_vars,
         workflow_run_id=workflow_run_id,

api/services/workflow/pipecat_engine.py

@@ -60,6 +60,7 @@ class PipecatEngine:
         *,
         task: Optional[PipelineTask] = None,
         llm: Optional["LLMService"] = None,
+        inference_llm: Optional["LLMService"] = None,
         context: Optional[LLMContext] = None,
         workflow: WorkflowGraph,
         call_context_vars: dict,
@@ -75,6 +76,12 @@ class PipecatEngine:
     ):
         self.task = task
         self.llm = llm
+        # LLM used for out-of-band inference (variable extraction, context
+        # summarization). Falls back to the pipeline LLM when not provided.
+        # In realtime mode the pipeline LLM is a speech-to-speech service
+        # that does not implement run_inference, so a separate text LLM
+        # must be passed in.
+        self.inference_llm = inference_llm or llm
         self.context = context
         self.workflow = workflow
         self._call_context_vars = call_context_vars

api/services/workflow/pipecat_engine_context_summarizer.py

@@ -63,7 +63,7 @@ class ContextSummarizationManager:
         orphaned tool calls from previous nodes) with a concise summary.
         """
         context = self._engine.context
-        llm = self._engine.llm
+        llm = self._engine.inference_llm
         current_node = self._engine._current_node
 
         try:

api/services/workflow/pipecat_engine_variable_extractor.py

@@ -203,12 +203,12 @@ class VariableExtractionManager:
         # current node's system prompt that build_chat_completion_params
         # would otherwise prepend.
         # ------------------------------------------------------------------
-        llm_response = await self._engine.llm.run_inference(
+        llm_response = await self._engine.inference_llm.run_inference(
             extraction_context, system_instruction=system_prompt
         )
 
         # Get model name for tracing
-        model_name = getattr(self._engine.llm, "model_name", "unknown")
+        model_name = getattr(self._engine.inference_llm, "model_name", "unknown")
 
         if ensure_tracing():
             tracer = trace.get_tracer("pipecat")
@@ -221,7 +221,7 @@ class VariableExtractionManager:
                 ]
                 add_llm_span_attributes(
                     span,
-                    service_name=self._engine.llm.__class__.__name__,
+                    service_name=self._engine.inference_llm.__class__.__name__,
                     model=model_name,
                     operation_name="llm-variable-extraction",
                     messages=tracing_messages,

docs/configurations/inference-providers.mdx

@@ -73,7 +73,7 @@ For example, if you only want to change the voice for a specific agent:
 You can also switch an individual agent to use a **Realtime** provider (such as Gemini Live) even if the global configuration uses standard LLM + TTS + STT. Toggle the **Realtime** switch in the Model Overrides tab, then configure the realtime provider, model, and voice.
 
 <Note>
-When an agent uses a Realtime provider, it replaces the separate LLM, TTS, and STT services with a single speech-to-speech model. The individual LLM/TTS/STT override tabs are hidden in this mode.
+When an agent uses a Realtime provider, it replaces the separate TTS and STT services with a single speech-to-speech model. An **LLM** is still required alongside the Realtime model — it's used for out-of-band tasks like variable extraction and QA analysis, which the realtime service does not handle. Context compaction is not applicable in Realtime mode and is ignored if enabled.
 </Note>
 
 ## Gemini 3.1 Live
@@ -119,5 +119,5 @@ To use Gemini 3.1 Live with Dograh, you need a Google Gemini API key. Follow the
 6. Select the language (currently `en` is supported).
 
 <Note>
-  When using a Realtime provider like Gemini Live, you do not need to configure separate LLM, TTS, and STT services — the realtime model handles all three.
+  When using a Realtime provider like Gemini Live, you do not need to configure separate TTS and STT services — the realtime model handles speech in and out. However, you **must** still configure an **LLM** under the LLM tab: it powers variable extraction and QA analysis, which the realtime service does not perform.
 </Note>

pipecat

@@ -1 +1 @@
-Subproject commit 49f1965d652fb5027968dcc24677bc83c5f905ab
+Subproject commit edefaad42b97e52a3ad5eef8d15115a5c6ba3b11

ui/.env.example

@@ -1,2 +1,3 @@
 BACKEND_URL=http://localhost:8000
+NEXT_PUBLIC_BACKEND_URL=http://localhost:8000
 NEXT_PUBLIC_NODE_ENV=development

ui/src/app/workflow/[workflowId]/components/WorkflowEditorHeader.tsx

@@ -385,7 +385,7 @@ export const WorkflowEditorHeader = ({
 
                 {/* GitHub star badge - desktop only */}
                 <div className="hidden md:block">
-                    <GitHubStarBadge className="border-[#3a3a3a] text-white" source="workflow_editor_header" />
+                    <GitHubStarBadge className="border-[#3a3a3a] bg-[#2a2a2a] text-white [&_span]:bg-transparent" source="workflow_editor_header" />
                 </div>
             </div>
         </div>

ui/src/app/workflow/[workflowId]/run/[runId]/components/UnifiedTranscript.tsx

@@ -84,7 +84,7 @@ export const UnifiedTranscript = ({
                             }
                             return (
                                 <TranscriptMessage
-                                    key={msg.id}
+                                    key={`${msg.id}-${index}`}
                                     message={msg}
                                     nextMessage={transcriptMessages[index + 1]}
                                 />

ui/src/app/workflow/[workflowId]/settings/page.tsx

@@ -582,7 +582,7 @@ function GeneralSection({
                     <div>
                         <h3 className="text-sm font-medium">Context Compaction</h3>
                         <p className="text-xs text-muted-foreground mt-0.5">
-                            Automatically summarize conversation context when transitioning between nodes.
+                            Automatically summarize conversation context when transitioning between nodes. Not applicable in Realtime mode — the speech-to-speech service manages its own conversation state and this setting is ignored.
                         </p>
                     </div>
                     <div className="flex items-center justify-between">

ui/src/components/ServiceConfigurationForm.tsx

@@ -52,6 +52,7 @@ const STANDARD_TABS: { key: ServiceSegment; label: string }[] = [
 
 const REALTIME_TABS: { key: ServiceSegment; label: string }[] = [
     { key: "realtime", label: "Realtime Model" },
+    { key: "llm", label: "LLM" },
     { key: "embeddings", label: "Embedding" },
 ];
 
@@ -63,6 +64,7 @@ const OVERRIDE_STANDARD_TABS: { key: ServiceSegment; label: string }[] = [
 
 const OVERRIDE_REALTIME_TABS: { key: ServiceSegment; label: string }[] = [
     { key: "realtime", label: "Realtime Model" },
+    { key: "llm", label: "LLM" },
 ];
 
 // Display names for Sarvam voices
@@ -407,7 +409,7 @@ export function ServiceConfigurationForm({
             if (mode === 'override') {
                 // Build model_overrides for enabled services only
                 const modelOverrides: Record<string, unknown> = {};
-                const services = isRealtime ? ["realtime"] : ["llm", "tts", "stt"];
+                const services = isRealtime ? ["realtime", "llm"] : ["llm", "tts", "stt"];
                 for (const svc of services) {
                     if (enabledOverrides[svc]) {
                         modelOverrides[svc] = buildServiceConfig(svc as ServiceSegment, data);
@@ -758,7 +760,7 @@ export function ServiceConfigurationForm({
                         Realtime Mode
                     </Label>
                     <p className="text-xs text-muted-foreground mt-0.5">
-                        Uses a single speech-to-speech model (no separate STT/TTS)
+                        Uses a single speech-to-speech model (no separate STT/TTS). An LLM is still required for variable extraction and QA.
                     </p>
                 </div>
                 <Switch