Always apply a timeout to summarization LLM calls

Even when summarization_timeout is explicitly set to None, use a
DEFAULT_SUMMARIZATION_TIMEOUT (120s) fallback so the LLM call can
never hang indefinitely. Applied in both LLMService and the dedicated
LLM path in LLMContextSummarizer.

src/pipecat/frames/frames.py

@@ -2020,7 +2020,7 @@ class LLMContextSummaryRequestFrame(ControlFrame):
         summarization_prompt: System prompt instructing the LLM how to generate
             the summary.
         summarization_timeout: Maximum time in seconds for the LLM to generate a
-            summary. None means no timeout.
+            summary. When None, a default timeout of 120s is applied.
     """
 
     request_id: str

src/pipecat/processors/aggregators/llm_context_summarizer.py

@@ -24,6 +24,7 @@ from pipecat.processors.aggregators.llm_context import LLMContext, LLMSpecificMe
 from pipecat.utils.asyncio.task_manager import BaseTaskManager
 from pipecat.utils.base_object import BaseObject
 from pipecat.utils.context.llm_context_summarization import (
+    DEFAULT_SUMMARIZATION_TIMEOUT,
     LLMContextSummarizationConfig,
     LLMContextSummarizationUtil,
 )
@@ -278,21 +279,20 @@ class LLMContextSummarizer(BaseObject):
             llm: The dedicated LLM service to use for summarization.
             frame: The summarization request frame.
         """
+        timeout = frame.summarization_timeout or DEFAULT_SUMMARIZATION_TIMEOUT
+
         try:
-            if frame.summarization_timeout:
-                summary, last_index = await asyncio.wait_for(
-                    llm._generate_summary(frame),
-                    timeout=frame.summarization_timeout,
-                )
-            else:
-                summary, last_index = await llm._generate_summary(frame)
+            summary, last_index = await asyncio.wait_for(
+                llm._generate_summary(frame),
+                timeout=timeout,
+            )
             result_frame = LLMContextSummaryResultFrame(
                 request_id=frame.request_id,
                 summary=summary,
                 last_summarized_index=last_index,
             )
         except asyncio.TimeoutError:
-            error = f"Context summarization timed out after {frame.summarization_timeout}s"
+            error = f"Context summarization timed out after {timeout}s"
             logger.error(f"{self}: {error}")
             result_frame = LLMContextSummaryResultFrame(
                 request_id=frame.request_id,

src/pipecat/services/llm_service.py

@@ -62,6 +62,7 @@ from pipecat.services.ai_service import AIService
 from pipecat.services.settings import LLMSettings
 from pipecat.turns.user_turn_completion_mixin import UserTurnCompletionLLMServiceMixin
 from pipecat.utils.context.llm_context_summarization import (
+    DEFAULT_SUMMARIZATION_TIMEOUT,
     LLMContextSummarizationUtil,
 )
 
@@ -436,18 +437,15 @@ class LLMService(UserTurnCompletionLLMServiceMixin, AIService):
         last_index = -1
         error = None
 
+        timeout = frame.summarization_timeout or DEFAULT_SUMMARIZATION_TIMEOUT
+
         try:
-            if frame.summarization_timeout:
-                summary, last_index = await asyncio.wait_for(
-                    self._generate_summary(frame),
-                    timeout=frame.summarization_timeout,
-                )
-            else:
-                summary, last_index = await self._generate_summary(frame)
-        except asyncio.TimeoutError:
-            await self.push_error(
-                error_msg=f"Context summarization timed out after {frame.summarization_timeout}s"
+            summary, last_index = await asyncio.wait_for(
+                self._generate_summary(frame),
+                timeout=timeout,
             )
+        except asyncio.TimeoutError:
+            await self.push_error(error_msg=f"Context summarization timed out after {timeout}s")
         except Exception as e:
             error = f"Error generating context summary: {e}"
             await self.push_error(error, exception=e)

src/pipecat/utils/context/llm_context_summarization.py

@@ -20,6 +20,9 @@ from loguru import logger
 
 from pipecat.processors.aggregators.llm_context import LLMContext, LLMSpecificMessage
 
+# Fallback timeout (seconds) used when summarization_timeout is None.
+DEFAULT_SUMMARIZATION_TIMEOUT = 120.0
+
 # Token estimation constants
 CHARS_PER_TOKEN = 4  # Industry-standard heuristic: 1 token ≈ 4 characters
 TOKEN_OVERHEAD_PER_MESSAGE = 10  # Estimated structural overhead per message
@@ -89,7 +92,6 @@ class LLMContextSummarizationConfig:
         summarization_timeout: Maximum time in seconds to wait for the LLM to
             generate a summary. If the call exceeds this timeout, summarization
             is aborted with an error and future summarizations are unblocked.
-            Set to None to disable the timeout.
     """
 
     max_context_tokens: int = 8000
@@ -99,7 +101,7 @@ class LLMContextSummarizationConfig:
     summarization_prompt: Optional[str] = None
     summary_message_template: str = "Conversation summary: {summary}"
     llm: Optional["LLMService"] = None
-    summarization_timeout: Optional[float] = 120.0
+    summarization_timeout: float = DEFAULT_SUMMARIZATION_TIMEOUT
 
     def __post_init__(self):
         """Validate configuration parameters."""