Add call_id to LLM events for correlating requests

When monitoring LLM events, consumers need to know which events belong
to the same API call. Before this change, there was no way to correlate
LLMCallStartedEvent, LLMStreamChunkEvent, and LLMCallCompletedEvent
belonging to the same request.

lib/crewai/src/crewai/events/types/llm_events.py

@@ -10,6 +10,7 @@ class LLMEventBase(BaseEvent):
     from_task: Any | None = None
     from_agent: Any | None = None
     model: str | None = None
+    call_id: str
 
     def __init__(self, **data: Any) -> None:
         if data.get("from_task"):

lib/crewai/src/crewai/llm.py

@@ -37,7 +37,7 @@ from crewai.events.types.tool_usage_events import (
     ToolUsageFinishedEvent,
     ToolUsageStartedEvent,
 )
-from crewai.llms.base_llm import BaseLLM
+from crewai.llms.base_llm import BaseLLM, get_current_call_id, llm_call_context
 from crewai.llms.constants import (
     ANTHROPIC_MODELS,
     AZURE_MODELS,
@@ -770,7 +770,7 @@ class LLM(BaseLLM):
                 chunk_content = None
                 response_id = None
 
-                if hasattr(chunk,'id'):
+                if hasattr(chunk, "id"):
                     response_id = chunk.id
 
                 # Safely extract content from various chunk formats
@@ -827,7 +827,7 @@ class LLM(BaseLLM):
                                         available_functions=available_functions,
                                         from_task=from_task,
                                         from_agent=from_agent,
-                                        response_id=response_id
+                                        response_id=response_id,
                                     )
 
                                     if result is not None:
@@ -849,7 +849,8 @@ class LLM(BaseLLM):
                             from_task=from_task,
                             from_agent=from_agent,
                             call_type=LLMCallType.LLM_CALL,
-                            response_id=response_id
+                            response_id=response_id,
+                            call_id=get_current_call_id(),
                         ),
                     )
             # --- 4) Fallback to non-streaming if no content received
@@ -1015,7 +1016,10 @@ class LLM(BaseLLM):
             crewai_event_bus.emit(
                 self,
                 event=LLMCallFailedEvent(
-                    error=str(e), from_task=from_task, from_agent=from_agent
+                    error=str(e),
+                    from_task=from_task,
+                    from_agent=from_agent,
+                    call_id=get_current_call_id(),
                 ),
             )
             raise Exception(f"Failed to get streaming response: {e!s}") from e
@@ -1048,7 +1052,8 @@ class LLM(BaseLLM):
                     from_task=from_task,
                     from_agent=from_agent,
                     call_type=LLMCallType.TOOL_CALL,
-                    response_id=response_id
+                    response_id=response_id,
+                    call_id=get_current_call_id(),
                 ),
             )
 
@@ -1476,7 +1481,8 @@ class LLM(BaseLLM):
                             chunk=chunk_content,
                             from_task=from_task,
                             from_agent=from_agent,
-                            response_id=response_id
+                            response_id=response_id,
+                            call_id=get_current_call_id(),
                         ),
                     )
 
@@ -1619,7 +1625,12 @@ class LLM(BaseLLM):
                 logging.error(f"Error executing function '{function_name}': {e}")
                 crewai_event_bus.emit(
                     self,
-                    event=LLMCallFailedEvent(error=f"Tool execution error: {e!s}"),
+                    event=LLMCallFailedEvent(
+                        error=f"Tool execution error: {e!s}",
+                        from_task=from_task,
+                        from_agent=from_agent,
+                        call_id=get_current_call_id(),
+                    ),
                 )
                 crewai_event_bus.emit(
                     self,
@@ -1669,6 +1680,7 @@ class LLM(BaseLLM):
             ValueError: If response format is not supported
             LLMContextLengthExceededError: If input exceeds model's context limit
         """
+        with llm_call_context() as call_id:
             crewai_event_bus.emit(
                 self,
                 event=LLMCallStartedEvent(
@@ -1679,6 +1691,7 @@ class LLM(BaseLLM):
                     from_task=from_task,
                     from_agent=from_agent,
                     model=self.model,
+                    call_id=call_id,
                 ),
             )
 
@@ -1748,9 +1761,13 @@ class LLM(BaseLLM):
                                 self.additional_params["additional_drop_params"], list
                             )
                         ):
-                        self.additional_params["additional_drop_params"].append("stop")
+                            self.additional_params["additional_drop_params"].append(
+                                "stop"
+                            )
                         else:
-                        self.additional_params = {"additional_drop_params": ["stop"]}
+                            self.additional_params = {
+                                "additional_drop_params": ["stop"]
+                            }
 
                         logging.info("Retrying LLM call without the unsupported 'stop'")
 
@@ -1767,7 +1784,10 @@ class LLM(BaseLLM):
                     crewai_event_bus.emit(
                         self,
                         event=LLMCallFailedEvent(
-                        error=str(e), from_task=from_task, from_agent=from_agent
+                            error=str(e),
+                            from_task=from_task,
+                            from_agent=from_agent,
+                            call_id=get_current_call_id(),
                         ),
                     )
                     raise
@@ -1808,6 +1828,7 @@ class LLM(BaseLLM):
             ValueError: If response format is not supported
             LLMContextLengthExceededError: If input exceeds model's context limit
         """
+        with llm_call_context() as call_id:
             crewai_event_bus.emit(
                 self,
                 event=LLMCallStartedEvent(
@@ -1818,6 +1839,7 @@ class LLM(BaseLLM):
                     from_task=from_task,
                     from_agent=from_agent,
                     model=self.model,
+                    call_id=call_id,
                 ),
             )
 
@@ -1875,9 +1897,13 @@ class LLM(BaseLLM):
                                 self.additional_params["additional_drop_params"], list
                             )
                         ):
-                        self.additional_params["additional_drop_params"].append("stop")
+                            self.additional_params["additional_drop_params"].append(
+                                "stop"
+                            )
                         else:
-                        self.additional_params = {"additional_drop_params": ["stop"]}
+                            self.additional_params = {
+                                "additional_drop_params": ["stop"]
+                            }
 
                         logging.info("Retrying LLM call without the unsupported 'stop'")
 
@@ -1894,7 +1920,10 @@ class LLM(BaseLLM):
                     crewai_event_bus.emit(
                         self,
                         event=LLMCallFailedEvent(
-                        error=str(e), from_task=from_task, from_agent=from_agent
+                            error=str(e),
+                            from_task=from_task,
+                            from_agent=from_agent,
+                            call_id=get_current_call_id(),
                         ),
                     )
                     raise
@@ -1925,6 +1954,7 @@ class LLM(BaseLLM):
                 from_task=from_task,
                 from_agent=from_agent,
                 model=self.model,
+                call_id=get_current_call_id(),
             ),
         )
 

lib/crewai/src/crewai/llms/base_llm.py

@@ -7,11 +7,15 @@ in CrewAI, including common functionality for native SDK implementations.
 from __future__ import annotations
 
 from abc import ABC, abstractmethod
+from collections.abc import Generator
+from contextlib import contextmanager
+import contextvars
 from datetime import datetime
 import json
 import logging
 import re
 from typing import TYPE_CHECKING, Any, Final
+import uuid
 
 from pydantic import BaseModel
 
@@ -50,6 +54,32 @@ DEFAULT_CONTEXT_WINDOW_SIZE: Final[int] = 4096
 DEFAULT_SUPPORTS_STOP_WORDS: Final[bool] = True
 _JSON_EXTRACTION_PATTERN: Final[re.Pattern[str]] = re.compile(r"\{.*}", re.DOTALL)
 
+_current_call_id: contextvars.ContextVar[str | None] = contextvars.ContextVar(
+    "_current_call_id", default=None
+)
+
+
+@contextmanager
+def llm_call_context() -> Generator[str, None, None]:
+    """Context manager that establishes an LLM call scope with a unique call_id."""
+    call_id = str(uuid.uuid4())
+    token = _current_call_id.set(call_id)
+    try:
+        yield call_id
+    finally:
+        _current_call_id.reset(token)
+
+
+def get_current_call_id() -> str:
+    """Get current call_id from context"""
+    call_id = _current_call_id.get()
+    if call_id is None:
+        logging.warning(
+            "LLM event emitted outside call context - generating fallback call_id"
+        )
+        return str(uuid.uuid4())
+    return call_id
+
 
 class BaseLLM(ABC):
     """Abstract base class for LLM implementations.
@@ -351,6 +381,7 @@ class BaseLLM(ABC):
                 from_task=from_task,
                 from_agent=from_agent,
                 model=self.model,
+                call_id=get_current_call_id(),
             ),
         )
 
@@ -374,6 +405,7 @@ class BaseLLM(ABC):
                 from_task=from_task,
                 from_agent=from_agent,
                 model=self.model,
+                call_id=get_current_call_id(),
             ),
         )
 
@@ -394,6 +426,7 @@ class BaseLLM(ABC):
                 from_task=from_task,
                 from_agent=from_agent,
                 model=self.model,
+                call_id=get_current_call_id(),
             ),
         )
 
@@ -404,7 +437,7 @@ class BaseLLM(ABC):
         from_agent: Agent | None = None,
         tool_call: dict[str, Any] | None = None,
         call_type: LLMCallType | None = None,
-        response_id: str | None = None
+        response_id: str | None = None,
     ) -> None:
         """Emit stream chunk event.
 
@@ -427,7 +460,8 @@ class BaseLLM(ABC):
                 from_task=from_task,
                 from_agent=from_agent,
                 call_type=call_type,
-                response_id=response_id
+                response_id=response_id,
+                call_id=get_current_call_id(),
             ),
         )
 

lib/crewai/src/crewai/llms/providers/anthropic/completion.py

@@ -9,7 +9,7 @@ from anthropic.types import ThinkingBlock
 from pydantic import BaseModel
 
 from crewai.events.types.llm_events import LLMCallType
-from crewai.llms.base_llm import BaseLLM
+from crewai.llms.base_llm import BaseLLM, llm_call_context
 from crewai.llms.hooks.transport import AsyncHTTPTransport, HTTPTransport
 from crewai.utilities.agent_utils import is_context_length_exceeded
 from crewai.utilities.exceptions.context_window_exceeding_exception import (
@@ -207,6 +207,7 @@ class AnthropicCompletion(BaseLLM):
         Returns:
             Chat completion response or tool call result
         """
+        with llm_call_context():
             try:
                 # Emit call started event
                 self._emit_call_started_event(
@@ -219,11 +220,13 @@ class AnthropicCompletion(BaseLLM):
                 )
 
                 # Format messages for Anthropic
-            formatted_messages, system_message = self._format_messages_for_anthropic(
+                formatted_messages, system_message = (
-                messages
+                    self._format_messages_for_anthropic(messages)
                 )
 
-            if not self._invoke_before_llm_call_hooks(formatted_messages, from_agent):
+                if not self._invoke_before_llm_call_hooks(
+                    formatted_messages, from_agent
+                ):
                     raise ValueError("LLM call blocked by before_llm_call hook")
 
                 # Prepare completion parameters
@@ -280,6 +283,7 @@ class AnthropicCompletion(BaseLLM):
         Returns:
             Chat completion response or tool call result
         """
+        with llm_call_context():
             try:
                 self._emit_call_started_event(
                     messages=messages,
@@ -290,8 +294,8 @@ class AnthropicCompletion(BaseLLM):
                     from_agent=from_agent,
                 )
 
-            formatted_messages, system_message = self._format_messages_for_anthropic(
+                formatted_messages, system_message = (
-                messages
+                    self._format_messages_for_anthropic(messages)
                 )
 
                 completion_params = self._prepare_completion_params(
@@ -712,7 +716,7 @@ class AnthropicCompletion(BaseLLM):
                         chunk=text_delta,
                         from_task=from_task,
                         from_agent=from_agent,
-                        response_id=response_id
+                        response_id=response_id,
                     )
 
                 if event.type == "content_block_start":
@@ -739,7 +743,7 @@ class AnthropicCompletion(BaseLLM):
                                 "index": block_index,
                             },
                             call_type=LLMCallType.TOOL_CALL,
-                            response_id=response_id
+                            response_id=response_id,
                         )
                 elif event.type == "content_block_delta":
                     if event.delta.type == "input_json_delta":
@@ -763,7 +767,7 @@ class AnthropicCompletion(BaseLLM):
                                     "index": block_index,
                                 },
                                 call_type=LLMCallType.TOOL_CALL,
-                                response_id=response_id
+                                response_id=response_id,
                             )
 
             final_message: Message = stream.get_final_message()
@@ -1133,7 +1137,7 @@ class AnthropicCompletion(BaseLLM):
                         chunk=text_delta,
                         from_task=from_task,
                         from_agent=from_agent,
-                        response_id=response_id
+                        response_id=response_id,
                     )
 
                 if event.type == "content_block_start":
@@ -1160,7 +1164,7 @@ class AnthropicCompletion(BaseLLM):
                                 "index": block_index,
                             },
                             call_type=LLMCallType.TOOL_CALL,
-                            response_id=response_id
+                            response_id=response_id,
                         )
                 elif event.type == "content_block_delta":
                     if event.delta.type == "input_json_delta":
@@ -1184,7 +1188,7 @@ class AnthropicCompletion(BaseLLM):
                                     "index": block_index,
                                 },
                                 call_type=LLMCallType.TOOL_CALL,
-                                response_id=response_id
+                                response_id=response_id,
                             )
 
             final_message: Message = await stream.get_final_message()

lib/crewai/src/crewai/llms/providers/azure/completion.py

@@ -43,7 +43,7 @@ try:
     )
 
     from crewai.events.types.llm_events import LLMCallType
-    from crewai.llms.base_llm import BaseLLM
+    from crewai.llms.base_llm import BaseLLM, llm_call_context
 
 except ImportError:
     raise ImportError(
@@ -288,6 +288,7 @@ class AzureCompletion(BaseLLM):
         Returns:
             Chat completion response or tool call result
         """
+        with llm_call_context():
             try:
                 # Emit call started event
                 self._emit_call_started_event(
@@ -302,7 +303,9 @@ class AzureCompletion(BaseLLM):
                 # Format messages for Azure
                 formatted_messages = self._format_messages_for_azure(messages)
 
-            if not self._invoke_before_llm_call_hooks(formatted_messages, from_agent):
+                if not self._invoke_before_llm_call_hooks(
+                    formatted_messages, from_agent
+                ):
                     raise ValueError("LLM call blocked by before_llm_call hook")
 
                 # Prepare completion parameters
@@ -355,6 +358,7 @@ class AzureCompletion(BaseLLM):
         Returns:
             Chat completion response or tool call result
         """
+        with llm_call_context():
             try:
                 self._emit_call_started_event(
                     messages=messages,
@@ -726,7 +730,7 @@ class AzureCompletion(BaseLLM):
         """
         if update.choices:
             choice = update.choices[0]
-            response_id = update.id if hasattr(update,"id") else None
+            response_id = update.id if hasattr(update, "id") else None
             if choice.delta and choice.delta.content:
                 content_delta = choice.delta.content
                 full_response += content_delta
@@ -734,7 +738,7 @@ class AzureCompletion(BaseLLM):
                     chunk=content_delta,
                     from_task=from_task,
                     from_agent=from_agent,
-                    response_id=response_id
+                    response_id=response_id,
                 )
 
             if choice.delta and choice.delta.tool_calls:
@@ -769,7 +773,7 @@ class AzureCompletion(BaseLLM):
                             "index": idx,
                         },
                         call_type=LLMCallType.TOOL_CALL,
-                        response_id=response_id
+                        response_id=response_id,
                     )
 
         return full_response

lib/crewai/src/crewai/llms/providers/bedrock/completion.py

@@ -11,7 +11,7 @@ from pydantic import BaseModel
 from typing_extensions import Required
 
 from crewai.events.types.llm_events import LLMCallType
-from crewai.llms.base_llm import BaseLLM
+from crewai.llms.base_llm import BaseLLM, llm_call_context
 from crewai.utilities.agent_utils import is_context_length_exceeded
 from crewai.utilities.exceptions.context_window_exceeding_exception import (
     LLMContextLengthExceededError,
@@ -299,6 +299,7 @@ class BedrockCompletion(BaseLLM):
         response_model: type[BaseModel] | None = None,
     ) -> str | Any:
         """Call AWS Bedrock Converse API."""
+        with llm_call_context():
             try:
                 # Emit call started event
                 self._emit_call_started_event(
@@ -315,7 +316,9 @@ class BedrockCompletion(BaseLLM):
                     messages
                 )
 
-            if not self._invoke_before_llm_call_hooks(formatted_messages, from_agent):
+                if not self._invoke_before_llm_call_hooks(
+                    formatted_messages, from_agent
+                ):
                     raise ValueError("LLM call blocked by before_llm_call hook")
 
                 # Prepare request body
@@ -354,7 +357,8 @@ class BedrockCompletion(BaseLLM):
                 # Add optional advanced features if configured
                 if self.guardrail_config:
                     guardrail_config: GuardrailConfigurationTypeDef = cast(
-                    "GuardrailConfigurationTypeDef", cast(object, self.guardrail_config)
+                        "GuardrailConfigurationTypeDef",
+                        cast(object, self.guardrail_config),
                     )
                     body["guardrailConfig"] = guardrail_config
 
@@ -431,6 +435,7 @@ class BedrockCompletion(BaseLLM):
                 'Install with: uv add "crewai[bedrock-async]"'
             )
 
+        with llm_call_context():
             try:
                 self._emit_call_started_event(
                     messages=messages,
@@ -478,7 +483,8 @@ class BedrockCompletion(BaseLLM):
 
                 if self.guardrail_config:
                     guardrail_config: GuardrailConfigurationTypeDef = cast(
-                    "GuardrailConfigurationTypeDef", cast(object, self.guardrail_config)
+                        "GuardrailConfigurationTypeDef",
+                        cast(object, self.guardrail_config),
                     )
                     body["guardrailConfig"] = guardrail_config
 
@@ -494,7 +500,11 @@ class BedrockCompletion(BaseLLM):
 
                 if self.stream:
                     return await self._ahandle_streaming_converse(
-                    formatted_messages, body, available_functions, from_task, from_agent
+                        formatted_messages,
+                        body,
+                        available_functions,
+                        from_task,
+                        from_agent,
                     )
 
                 return await self._ahandle_converse(
@@ -805,7 +815,7 @@ class BedrockCompletion(BaseLLM):
                                         "index": tool_use_index,
                                     },
                                     call_type=LLMCallType.TOOL_CALL,
-                                    response_id=response_id
+                                    response_id=response_id,
                                 )
                     elif "contentBlockStop" in event:
                         logging.debug("Content block stopped in stream")
@@ -1174,7 +1184,7 @@ class BedrockCompletion(BaseLLM):
                                 chunk=text_chunk,
                                 from_task=from_task,
                                 from_agent=from_agent,
-                                response_id=response_id
+                                response_id=response_id,
                             )
                         elif "toolUse" in delta and current_tool_use:
                             tool_input = delta["toolUse"].get("input", "")

lib/crewai/src/crewai/llms/providers/gemini/completion.py

@@ -10,7 +10,7 @@ from typing import TYPE_CHECKING, Any, Literal, cast
 from pydantic import BaseModel
 
 from crewai.events.types.llm_events import LLMCallType
-from crewai.llms.base_llm import BaseLLM
+from crewai.llms.base_llm import BaseLLM, llm_call_context
 from crewai.utilities.agent_utils import is_context_length_exceeded
 from crewai.utilities.exceptions.context_window_exceeding_exception import (
     LLMContextLengthExceededError,
@@ -282,6 +282,7 @@ class GeminiCompletion(BaseLLM):
         Returns:
             Chat completion response or tool call result
         """
+        with llm_call_context():
             try:
                 self._emit_call_started_event(
                     messages=messages,
@@ -293,13 +294,15 @@ class GeminiCompletion(BaseLLM):
                 )
                 self.tools = tools
 
-            formatted_content, system_instruction = self._format_messages_for_gemini(
+                formatted_content, system_instruction = (
-                messages
+                    self._format_messages_for_gemini(messages)
                 )
 
                 messages_for_hooks = self._convert_contents_to_dict(formatted_content)
 
-            if not self._invoke_before_llm_call_hooks(messages_for_hooks, from_agent):
+                if not self._invoke_before_llm_call_hooks(
+                    messages_for_hooks, from_agent
+                ):
                     raise ValueError("LLM call blocked by before_llm_call hook")
 
                 config = self._prepare_generation_config(
@@ -364,6 +367,7 @@ class GeminiCompletion(BaseLLM):
         Returns:
             Chat completion response or tool call result
         """
+        with llm_call_context():
             try:
                 self._emit_call_started_event(
                     messages=messages,
@@ -375,8 +379,8 @@ class GeminiCompletion(BaseLLM):
                 )
                 self.tools = tools
 
-            formatted_content, system_instruction = self._format_messages_for_gemini(
+                formatted_content, system_instruction = (
-                messages
+                    self._format_messages_for_gemini(messages)
                 )
 
                 config = self._prepare_generation_config(
@@ -790,7 +794,7 @@ class GeminiCompletion(BaseLLM):
         Returns:
             Tuple of (updated full_response, updated function_calls, updated usage_data)
         """
-        response_id=chunk.response_id if hasattr(chunk,"response_id") else None
+        response_id = chunk.response_id if hasattr(chunk, "response_id") else None
         if chunk.usage_metadata:
             usage_data = self._extract_token_usage(chunk)
 
@@ -800,7 +804,7 @@ class GeminiCompletion(BaseLLM):
                 chunk=chunk.text,
                 from_task=from_task,
                 from_agent=from_agent,
-                response_id=response_id
+                response_id=response_id,
             )
 
         if chunk.candidates:
@@ -837,7 +841,7 @@ class GeminiCompletion(BaseLLM):
                                 "index": call_index,
                             },
                             call_type=LLMCallType.TOOL_CALL,
-                            response_id=response_id
+                            response_id=response_id,
                         )
 
         return full_response, function_calls, usage_data

lib/crewai/src/crewai/llms/providers/openai/completion.py

@@ -17,7 +17,7 @@ from openai.types.responses import Response
 from pydantic import BaseModel
 
 from crewai.events.types.llm_events import LLMCallType
-from crewai.llms.base_llm import BaseLLM
+from crewai.llms.base_llm import BaseLLM, llm_call_context
 from crewai.llms.hooks.transport import AsyncHTTPTransport, HTTPTransport
 from crewai.utilities.agent_utils import is_context_length_exceeded
 from crewai.utilities.exceptions.context_window_exceeding_exception import (
@@ -382,6 +382,7 @@ class OpenAICompletion(BaseLLM):
         Returns:
             Completion response or tool call result.
         """
+        with llm_call_context():
             try:
                 self._emit_call_started_event(
                     messages=messages,
@@ -394,7 +395,9 @@ class OpenAICompletion(BaseLLM):
 
                 formatted_messages = self._format_messages(messages)
 
-            if not self._invoke_before_llm_call_hooks(formatted_messages, from_agent):
+                if not self._invoke_before_llm_call_hooks(
+                    formatted_messages, from_agent
+                ):
                     raise ValueError("LLM call blocked by before_llm_call hook")
 
                 if self.api == "responses":
@@ -479,6 +482,7 @@ class OpenAICompletion(BaseLLM):
         Returns:
             Completion response or tool call result.
         """
+        with llm_call_context():
             try:
                 self._emit_call_started_event(
                     messages=messages,
@@ -1060,7 +1064,7 @@ class OpenAICompletion(BaseLLM):
                     chunk=delta_text,
                     from_task=from_task,
                     from_agent=from_agent,
-                    response_id=response_id_stream
+                    response_id=response_id_stream,
                 )
 
             elif event.type == "response.function_call_arguments.delta":
@@ -1709,7 +1713,7 @@ class OpenAICompletion(BaseLLM):
                 **parse_params, response_format=response_model
             ) as stream:
                 for chunk in stream:
-                    response_id_stream=chunk.id if hasattr(chunk,"id") else None
+                    response_id_stream = chunk.id if hasattr(chunk, "id") else None
 
                     if chunk.type == "content.delta":
                         delta_content = chunk.delta
@@ -1718,7 +1722,7 @@ class OpenAICompletion(BaseLLM):
                                 chunk=delta_content,
                                 from_task=from_task,
                                 from_agent=from_agent,
-                                response_id=response_id_stream
+                                response_id=response_id_stream,
                             )
 
                 final_completion = stream.get_final_completion()
@@ -1748,7 +1752,9 @@ class OpenAICompletion(BaseLLM):
         usage_data = {"total_tokens": 0}
 
         for completion_chunk in completion_stream:
-            response_id_stream=completion_chunk.id if hasattr(completion_chunk,"id") else None
+            response_id_stream = (
+                completion_chunk.id if hasattr(completion_chunk, "id") else None
+            )
 
             if hasattr(completion_chunk, "usage") and completion_chunk.usage:
                 usage_data = self._extract_openai_token_usage(completion_chunk)
@@ -1766,7 +1772,7 @@ class OpenAICompletion(BaseLLM):
                     chunk=chunk_delta.content,
                     from_task=from_task,
                     from_agent=from_agent,
-                    response_id=response_id_stream
+                    response_id=response_id_stream,
                 )
 
             if chunk_delta.tool_calls:
@@ -1805,7 +1811,7 @@ class OpenAICompletion(BaseLLM):
                             "index": tool_calls[tool_index]["index"],
                         },
                         call_type=LLMCallType.TOOL_CALL,
-                        response_id=response_id_stream
+                        response_id=response_id_stream,
                     )
 
         self._track_token_usage_internal(usage_data)
@@ -2017,7 +2023,7 @@ class OpenAICompletion(BaseLLM):
             accumulated_content = ""
             usage_data = {"total_tokens": 0}
             async for chunk in completion_stream:
-                response_id_stream=chunk.id if hasattr(chunk,"id") else None
+                response_id_stream = chunk.id if hasattr(chunk, "id") else None
 
                 if hasattr(chunk, "usage") and chunk.usage:
                     usage_data = self._extract_openai_token_usage(chunk)
@@ -2035,7 +2041,7 @@ class OpenAICompletion(BaseLLM):
                         chunk=delta.content,
                         from_task=from_task,
                         from_agent=from_agent,
-                        response_id=response_id_stream
+                        response_id=response_id_stream,
                     )
 
             self._track_token_usage_internal(usage_data)
@@ -2071,7 +2077,7 @@ class OpenAICompletion(BaseLLM):
         usage_data = {"total_tokens": 0}
 
         async for chunk in stream:
-            response_id_stream=chunk.id if hasattr(chunk,"id") else None
+            response_id_stream = chunk.id if hasattr(chunk, "id") else None
 
             if hasattr(chunk, "usage") and chunk.usage:
                 usage_data = self._extract_openai_token_usage(chunk)
@@ -2089,7 +2095,7 @@ class OpenAICompletion(BaseLLM):
                     chunk=chunk_delta.content,
                     from_task=from_task,
                     from_agent=from_agent,
-                    response_id=response_id_stream
+                    response_id=response_id_stream,
                 )
 
             if chunk_delta.tool_calls:
@@ -2128,7 +2134,7 @@ class OpenAICompletion(BaseLLM):
                             "index": tool_calls[tool_index]["index"],
                         },
                         call_type=LLMCallType.TOOL_CALL,
-                        response_id=response_id_stream
+                        response_id=response_id_stream,
                     )
 
         self._track_token_usage_internal(usage_data)

lib/crewai/tests/cassettes/utilities/test_llm_call_events_share_call_id.yaml

@@ -0,0 +1,108 @@
+interactions:
+- request:
+    body: '{"messages":[{"role":"user","content":"Say hi"}],"model":"gpt-4o-mini"}'
+    headers:
+      User-Agent:
+      - X-USER-AGENT-XXX
+      accept:
+      - application/json
+      accept-encoding:
+      - ACCEPT-ENCODING-XXX
+      authorization:
+      - AUTHORIZATION-XXX
+      connection:
+      - keep-alive
+      content-length:
+      - '71'
+      content-type:
+      - application/json
+      host:
+      - api.openai.com
+      x-stainless-arch:
+      - X-STAINLESS-ARCH-XXX
+      x-stainless-async:
+      - 'false'
+      x-stainless-lang:
+      - python
+      x-stainless-os:
+      - X-STAINLESS-OS-XXX
+      x-stainless-package-version:
+      - 1.83.0
+      x-stainless-read-timeout:
+      - X-STAINLESS-READ-TIMEOUT-XXX
+      x-stainless-retry-count:
+      - '0'
+      x-stainless-runtime:
+      - CPython
+      x-stainless-runtime-version:
+      - 3.13.0
+    method: POST
+    uri: https://api.openai.com/v1/chat/completions
+  response:
+    body:
+      string: "{\n  \"id\": \"chatcmpl-D2HpUSxS5LeHwDTELElWlC5CDMzmr\",\n  \"object\":
+        \"chat.completion\",\n  \"created\": 1769437564,\n  \"model\": \"gpt-4o-mini-2024-07-18\",\n
+        \ \"choices\": [\n    {\n      \"index\": 0,\n      \"message\": {\n        \"role\":
+        \"assistant\",\n        \"content\": \"Hi there! How can I assist you today?\",\n
+        \       \"refusal\": null,\n        \"annotations\": []\n      },\n      \"logprobs\":
+        null,\n      \"finish_reason\": \"stop\"\n    }\n  ],\n  \"usage\": {\n    \"prompt_tokens\":
+        9,\n    \"completion_tokens\": 10,\n    \"total_tokens\": 19,\n    \"prompt_tokens_details\":
+        {\n      \"cached_tokens\": 0,\n      \"audio_tokens\": 0\n    },\n    \"completion_tokens_details\":
+        {\n      \"reasoning_tokens\": 0,\n      \"audio_tokens\": 0,\n      \"accepted_prediction_tokens\":
+        0,\n      \"rejected_prediction_tokens\": 0\n    }\n  },\n  \"service_tier\":
+        \"default\",\n  \"system_fingerprint\": \"fp_29330a9688\"\n}\n"
+    headers:
+      CF-RAY:
+      - CF-RAY-XXX
+      Connection:
+      - keep-alive
+      Content-Type:
+      - application/json
+      Date:
+      - Mon, 26 Jan 2026 14:26:05 GMT
+      Server:
+      - cloudflare
+      Set-Cookie:
+      - SET-COOKIE-XXX
+      Strict-Transport-Security:
+      - STS-XXX
+      Transfer-Encoding:
+      - chunked
+      X-Content-Type-Options:
+      - X-CONTENT-TYPE-XXX
+      access-control-expose-headers:
+      - ACCESS-CONTROL-XXX
+      alt-svc:
+      - h3=":443"; ma=86400
+      cf-cache-status:
+      - DYNAMIC
+      openai-organization:
+      - OPENAI-ORG-XXX
+      openai-processing-ms:
+      - '460'
+      openai-project:
+      - OPENAI-PROJECT-XXX
+      openai-version:
+      - '2020-10-01'
+      x-envoy-upstream-service-time:
+      - '477'
+      x-openai-proxy-wasm:
+      - v0.1
+      x-ratelimit-limit-requests:
+      - X-RATELIMIT-LIMIT-REQUESTS-XXX
+      x-ratelimit-limit-tokens:
+      - X-RATELIMIT-LIMIT-TOKENS-XXX
+      x-ratelimit-remaining-requests:
+      - X-RATELIMIT-REMAINING-REQUESTS-XXX
+      x-ratelimit-remaining-tokens:
+      - X-RATELIMIT-REMAINING-TOKENS-XXX
+      x-ratelimit-reset-requests:
+      - X-RATELIMIT-RESET-REQUESTS-XXX
+      x-ratelimit-reset-tokens:
+      - X-RATELIMIT-RESET-TOKENS-XXX
+      x-request-id:
+      - X-REQUEST-ID-XXX
+    status:
+      code: 200
+      message: OK
+version: 1

lib/crewai/tests/cassettes/utilities/test_separate_llm_calls_have_different_call_ids.yaml

@@ -0,0 +1,215 @@
+interactions:
+- request:
+    body: '{"messages":[{"role":"user","content":"Say hi"}],"model":"gpt-4o-mini"}'
+    headers:
+      User-Agent:
+      - X-USER-AGENT-XXX
+      accept:
+      - application/json
+      accept-encoding:
+      - ACCEPT-ENCODING-XXX
+      authorization:
+      - AUTHORIZATION-XXX
+      connection:
+      - keep-alive
+      content-length:
+      - '71'
+      content-type:
+      - application/json
+      host:
+      - api.openai.com
+      x-stainless-arch:
+      - X-STAINLESS-ARCH-XXX
+      x-stainless-async:
+      - 'false'
+      x-stainless-lang:
+      - python
+      x-stainless-os:
+      - X-STAINLESS-OS-XXX
+      x-stainless-package-version:
+      - 1.83.0
+      x-stainless-read-timeout:
+      - X-STAINLESS-READ-TIMEOUT-XXX
+      x-stainless-retry-count:
+      - '0'
+      x-stainless-runtime:
+      - CPython
+      x-stainless-runtime-version:
+      - 3.13.0
+    method: POST
+    uri: https://api.openai.com/v1/chat/completions
+  response:
+    body:
+      string: "{\n  \"id\": \"chatcmpl-D2HpStmyOpe9DrthWBlDdMZfVMJ1u\",\n  \"object\":
+        \"chat.completion\",\n  \"created\": 1769437562,\n  \"model\": \"gpt-4o-mini-2024-07-18\",\n
+        \ \"choices\": [\n    {\n      \"index\": 0,\n      \"message\": {\n        \"role\":
+        \"assistant\",\n        \"content\": \"Hi! How can I assist you today?\",\n
+        \       \"refusal\": null,\n        \"annotations\": []\n      },\n      \"logprobs\":
+        null,\n      \"finish_reason\": \"stop\"\n    }\n  ],\n  \"usage\": {\n    \"prompt_tokens\":
+        9,\n    \"completion_tokens\": 9,\n    \"total_tokens\": 18,\n    \"prompt_tokens_details\":
+        {\n      \"cached_tokens\": 0,\n      \"audio_tokens\": 0\n    },\n    \"completion_tokens_details\":
+        {\n      \"reasoning_tokens\": 0,\n      \"audio_tokens\": 0,\n      \"accepted_prediction_tokens\":
+        0,\n      \"rejected_prediction_tokens\": 0\n    }\n  },\n  \"service_tier\":
+        \"default\",\n  \"system_fingerprint\": \"fp_29330a9688\"\n}\n"
+    headers:
+      CF-RAY:
+      - CF-RAY-XXX
+      Connection:
+      - keep-alive
+      Content-Type:
+      - application/json
+      Date:
+      - Mon, 26 Jan 2026 14:26:02 GMT
+      Server:
+      - cloudflare
+      Set-Cookie:
+      - SET-COOKIE-XXX
+      Strict-Transport-Security:
+      - STS-XXX
+      Transfer-Encoding:
+      - chunked
+      X-Content-Type-Options:
+      - X-CONTENT-TYPE-XXX
+      access-control-expose-headers:
+      - ACCESS-CONTROL-XXX
+      alt-svc:
+      - h3=":443"; ma=86400
+      cf-cache-status:
+      - DYNAMIC
+      openai-organization:
+      - OPENAI-ORG-XXX
+      openai-processing-ms:
+      - '415'
+      openai-project:
+      - OPENAI-PROJECT-XXX
+      openai-version:
+      - '2020-10-01'
+      x-envoy-upstream-service-time:
+      - '434'
+      x-openai-proxy-wasm:
+      - v0.1
+      x-ratelimit-limit-requests:
+      - X-RATELIMIT-LIMIT-REQUESTS-XXX
+      x-ratelimit-limit-tokens:
+      - X-RATELIMIT-LIMIT-TOKENS-XXX
+      x-ratelimit-remaining-requests:
+      - X-RATELIMIT-REMAINING-REQUESTS-XXX
+      x-ratelimit-remaining-tokens:
+      - X-RATELIMIT-REMAINING-TOKENS-XXX
+      x-ratelimit-reset-requests:
+      - X-RATELIMIT-RESET-REQUESTS-XXX
+      x-ratelimit-reset-tokens:
+      - X-RATELIMIT-RESET-TOKENS-XXX
+      x-request-id:
+      - X-REQUEST-ID-XXX
+    status:
+      code: 200
+      message: OK
+- request:
+    body: '{"messages":[{"role":"user","content":"Say bye"}],"model":"gpt-4o-mini"}'
+    headers:
+      User-Agent:
+      - X-USER-AGENT-XXX
+      accept:
+      - application/json
+      accept-encoding:
+      - ACCEPT-ENCODING-XXX
+      authorization:
+      - AUTHORIZATION-XXX
+      connection:
+      - keep-alive
+      content-length:
+      - '72'
+      content-type:
+      - application/json
+      cookie:
+      - COOKIE-XXX
+      host:
+      - api.openai.com
+      x-stainless-arch:
+      - X-STAINLESS-ARCH-XXX
+      x-stainless-async:
+      - 'false'
+      x-stainless-lang:
+      - python
+      x-stainless-os:
+      - X-STAINLESS-OS-XXX
+      x-stainless-package-version:
+      - 1.83.0
+      x-stainless-read-timeout:
+      - X-STAINLESS-READ-TIMEOUT-XXX
+      x-stainless-retry-count:
+      - '0'
+      x-stainless-runtime:
+      - CPython
+      x-stainless-runtime-version:
+      - 3.13.0
+    method: POST
+    uri: https://api.openai.com/v1/chat/completions
+  response:
+    body:
+      string: "{\n  \"id\": \"chatcmpl-D2HpS1DP0Xd3tmWt5PBincVrdU7yw\",\n  \"object\":
+        \"chat.completion\",\n  \"created\": 1769437562,\n  \"model\": \"gpt-4o-mini-2024-07-18\",\n
+        \ \"choices\": [\n    {\n      \"index\": 0,\n      \"message\": {\n        \"role\":
+        \"assistant\",\n        \"content\": \"Goodbye! If you have more questions
+        in the future, feel free to reach out. Have a great day!\",\n        \"refusal\":
+        null,\n        \"annotations\": []\n      },\n      \"logprobs\": null,\n
+        \     \"finish_reason\": \"stop\"\n    }\n  ],\n  \"usage\": {\n    \"prompt_tokens\":
+        9,\n    \"completion_tokens\": 23,\n    \"total_tokens\": 32,\n    \"prompt_tokens_details\":
+        {\n      \"cached_tokens\": 0,\n      \"audio_tokens\": 0\n    },\n    \"completion_tokens_details\":
+        {\n      \"reasoning_tokens\": 0,\n      \"audio_tokens\": 0,\n      \"accepted_prediction_tokens\":
+        0,\n      \"rejected_prediction_tokens\": 0\n    }\n  },\n  \"service_tier\":
+        \"default\",\n  \"system_fingerprint\": \"fp_29330a9688\"\n}\n"
+    headers:
+      CF-RAY:
+      - CF-RAY-XXX
+      Connection:
+      - keep-alive
+      Content-Type:
+      - application/json
+      Date:
+      - Mon, 26 Jan 2026 14:26:03 GMT
+      Server:
+      - cloudflare
+      Strict-Transport-Security:
+      - STS-XXX
+      Transfer-Encoding:
+      - chunked
+      X-Content-Type-Options:
+      - X-CONTENT-TYPE-XXX
+      access-control-expose-headers:
+      - ACCESS-CONTROL-XXX
+      alt-svc:
+      - h3=":443"; ma=86400
+      cf-cache-status:
+      - DYNAMIC
+      openai-organization:
+      - OPENAI-ORG-XXX
+      openai-processing-ms:
+      - '964'
+      openai-project:
+      - OPENAI-PROJECT-XXX
+      openai-version:
+      - '2020-10-01'
+      x-envoy-upstream-service-time:
+      - '979'
+      x-openai-proxy-wasm:
+      - v0.1
+      x-ratelimit-limit-requests:
+      - X-RATELIMIT-LIMIT-REQUESTS-XXX
+      x-ratelimit-limit-tokens:
+      - X-RATELIMIT-LIMIT-TOKENS-XXX
+      x-ratelimit-remaining-requests:
+      - X-RATELIMIT-REMAINING-REQUESTS-XXX
+      x-ratelimit-remaining-tokens:
+      - X-RATELIMIT-REMAINING-TOKENS-XXX
+      x-ratelimit-reset-requests:
+      - X-RATELIMIT-RESET-REQUESTS-XXX
+      x-ratelimit-reset-tokens:
+      - X-RATELIMIT-RESET-TOKENS-XXX
+      x-request-id:
+      - X-REQUEST-ID-XXX
+    status:
+      code: 200
+      message: OK
+version: 1

lib/crewai/tests/cassettes/utilities/test_streaming_chunks_share_call_id_with_call.yaml

@@ -0,0 +1,143 @@
+interactions:
+- request:
+    body: '{"messages":[{"role":"user","content":"Say hi"}],"model":"gpt-4o-mini","stream":true,"stream_options":{"include_usage":true}}'
+    headers:
+      User-Agent:
+      - X-USER-AGENT-XXX
+      accept:
+      - application/json
+      accept-encoding:
+      - ACCEPT-ENCODING-XXX
+      authorization:
+      - AUTHORIZATION-XXX
+      connection:
+      - keep-alive
+      content-length:
+      - '125'
+      content-type:
+      - application/json
+      host:
+      - api.openai.com
+      x-stainless-arch:
+      - X-STAINLESS-ARCH-XXX
+      x-stainless-async:
+      - 'false'
+      x-stainless-lang:
+      - python
+      x-stainless-os:
+      - X-STAINLESS-OS-XXX
+      x-stainless-package-version:
+      - 1.83.0
+      x-stainless-read-timeout:
+      - X-STAINLESS-READ-TIMEOUT-XXX
+      x-stainless-retry-count:
+      - '0'
+      x-stainless-runtime:
+      - CPython
+      x-stainless-runtime-version:
+      - 3.13.0
+    method: POST
+    uri: https://api.openai.com/v1/chat/completions
+  response:
+    body:
+      string: 'data: {"id":"chatcmpl-D2HpUGTvIFKBsR9Xd6XRT4AuFXzbz","object":"chat.completion.chunk","created":1769437564,"model":"gpt-4o-mini-2024-07-18","service_tier":"default","system_fingerprint":"fp_29330a9688","choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"rVIyGQF2E"}
+
+
+        data: {"id":"chatcmpl-D2HpUGTvIFKBsR9Xd6XRT4AuFXzbz","object":"chat.completion.chunk","created":1769437564,"model":"gpt-4o-mini-2024-07-18","service_tier":"default","system_fingerprint":"fp_29330a9688","choices":[{"index":0,"delta":{"content":"Hi"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"ZGVqV7ZDm"}
+
+
+        data: {"id":"chatcmpl-D2HpUGTvIFKBsR9Xd6XRT4AuFXzbz","object":"chat.completion.chunk","created":1769437564,"model":"gpt-4o-mini-2024-07-18","service_tier":"default","system_fingerprint":"fp_29330a9688","choices":[{"index":0,"delta":{"content":"!"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"vnfm7IxlIB"}
+
+
+        data: {"id":"chatcmpl-D2HpUGTvIFKBsR9Xd6XRT4AuFXzbz","object":"chat.completion.chunk","created":1769437564,"model":"gpt-4o-mini-2024-07-18","service_tier":"default","system_fingerprint":"fp_29330a9688","choices":[{"index":0,"delta":{"content":"
+        How"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"o8F35ZZ"}
+
+
+        data: {"id":"chatcmpl-D2HpUGTvIFKBsR9Xd6XRT4AuFXzbz","object":"chat.completion.chunk","created":1769437564,"model":"gpt-4o-mini-2024-07-18","service_tier":"default","system_fingerprint":"fp_29330a9688","choices":[{"index":0,"delta":{"content":"
+        can"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"kiBzGe3"}
+
+
+        data: {"id":"chatcmpl-D2HpUGTvIFKBsR9Xd6XRT4AuFXzbz","object":"chat.completion.chunk","created":1769437564,"model":"gpt-4o-mini-2024-07-18","service_tier":"default","system_fingerprint":"fp_29330a9688","choices":[{"index":0,"delta":{"content":"
+        I"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"cbGT2RWgx"}
+
+
+        data: {"id":"chatcmpl-D2HpUGTvIFKBsR9Xd6XRT4AuFXzbz","object":"chat.completion.chunk","created":1769437564,"model":"gpt-4o-mini-2024-07-18","service_tier":"default","system_fingerprint":"fp_29330a9688","choices":[{"index":0,"delta":{"content":"
+        assist"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"DtxR"}
+
+
+        data: {"id":"chatcmpl-D2HpUGTvIFKBsR9Xd6XRT4AuFXzbz","object":"chat.completion.chunk","created":1769437564,"model":"gpt-4o-mini-2024-07-18","service_tier":"default","system_fingerprint":"fp_29330a9688","choices":[{"index":0,"delta":{"content":"
+        you"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"6y6Co8J"}
+
+
+        data: {"id":"chatcmpl-D2HpUGTvIFKBsR9Xd6XRT4AuFXzbz","object":"chat.completion.chunk","created":1769437564,"model":"gpt-4o-mini-2024-07-18","service_tier":"default","system_fingerprint":"fp_29330a9688","choices":[{"index":0,"delta":{"content":"
+        today"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"SZOmm"}
+
+
+        data: {"id":"chatcmpl-D2HpUGTvIFKBsR9Xd6XRT4AuFXzbz","object":"chat.completion.chunk","created":1769437564,"model":"gpt-4o-mini-2024-07-18","service_tier":"default","system_fingerprint":"fp_29330a9688","choices":[{"index":0,"delta":{"content":"?"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"s9Bc0HqlPg"}
+
+
+        data: {"id":"chatcmpl-D2HpUGTvIFKBsR9Xd6XRT4AuFXzbz","object":"chat.completion.chunk","created":1769437564,"model":"gpt-4o-mini-2024-07-18","service_tier":"default","system_fingerprint":"fp_29330a9688","choices":[{"index":0,"delta":{},"logprobs":null,"finish_reason":"stop"}],"usage":null,"obfuscation":"u9aar"}
+
+
+        data: {"id":"chatcmpl-D2HpUGTvIFKBsR9Xd6XRT4AuFXzbz","object":"chat.completion.chunk","created":1769437564,"model":"gpt-4o-mini-2024-07-18","service_tier":"default","system_fingerprint":"fp_29330a9688","choices":[],"usage":{"prompt_tokens":9,"completion_tokens":9,"total_tokens":18,"prompt_tokens_details":{"cached_tokens":0,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"5hudm8ySqh39"}
+
+
+        data: [DONE]
+
+
+        '
+    headers:
+      CF-RAY:
+      - CF-RAY-XXX
+      Connection:
+      - keep-alive
+      Content-Type:
+      - text/event-stream; charset=utf-8
+      Date:
+      - Mon, 26 Jan 2026 14:26:04 GMT
+      Server:
+      - cloudflare
+      Set-Cookie:
+      - SET-COOKIE-XXX
+      Strict-Transport-Security:
+      - STS-XXX
+      Transfer-Encoding:
+      - chunked
+      X-Content-Type-Options:
+      - X-CONTENT-TYPE-XXX
+      access-control-expose-headers:
+      - ACCESS-CONTROL-XXX
+      alt-svc:
+      - h3=":443"; ma=86400
+      cf-cache-status:
+      - DYNAMIC
+      openai-organization:
+      - OPENAI-ORG-XXX
+      openai-processing-ms:
+      - '260'
+      openai-project:
+      - OPENAI-PROJECT-XXX
+      openai-version:
+      - '2020-10-01'
+      x-envoy-upstream-service-time:
+      - '275'
+      x-openai-proxy-wasm:
+      - v0.1
+      x-ratelimit-limit-requests:
+      - X-RATELIMIT-LIMIT-REQUESTS-XXX
+      x-ratelimit-limit-tokens:
+      - X-RATELIMIT-LIMIT-TOKENS-XXX
+      x-ratelimit-remaining-requests:
+      - X-RATELIMIT-REMAINING-REQUESTS-XXX
+      x-ratelimit-remaining-tokens:
+      - X-RATELIMIT-REMAINING-TOKENS-XXX
+      x-ratelimit-reset-requests:
+      - X-RATELIMIT-RESET-REQUESTS-XXX
+      x-ratelimit-reset-tokens:
+      - X-RATELIMIT-RESET-TOKENS-XXX
+      x-request-id:
+      - X-REQUEST-ID-XXX
+    status:
+      code: 200
+      message: OK
+version: 1

lib/crewai/tests/test_streaming.py

@@ -217,6 +217,7 @@ class TestCrewKickoffStreaming:
                     LLMStreamChunkEvent(
                         type="llm_stream_chunk",
                         chunk="Hello ",
+                        call_id="test-call-id",
                     ),
                 )
                 crewai_event_bus.emit(
@@ -224,6 +225,7 @@ class TestCrewKickoffStreaming:
                     LLMStreamChunkEvent(
                         type="llm_stream_chunk",
                         chunk="World!",
+                        call_id="test-call-id",
                     ),
                 )
                 return mock_output
@@ -284,6 +286,7 @@ class TestCrewKickoffStreaming:
                     LLMStreamChunkEvent(
                         type="llm_stream_chunk",
                         chunk="",
+                        call_id="test-call-id",
                         tool_call=ToolCall(
                             id="call-123",
                             function=FunctionCall(
@@ -364,6 +367,7 @@ class TestCrewKickoffStreamingAsync:
                 LLMStreamChunkEvent(
                     type="llm_stream_chunk",
                     chunk="Async ",
+                    call_id="test-call-id",
                 ),
             )
             crewai_event_bus.emit(
@@ -371,6 +375,7 @@ class TestCrewKickoffStreamingAsync:
                 LLMStreamChunkEvent(
                     type="llm_stream_chunk",
                     chunk="Stream!",
+                    call_id="test-call-id",
                 ),
             )
             return mock_output
@@ -451,6 +456,7 @@ class TestFlowKickoffStreaming:
                     LLMStreamChunkEvent(
                         type="llm_stream_chunk",
                         chunk="Flow ",
+                        call_id="test-call-id",
                     ),
                 )
                 crewai_event_bus.emit(
@@ -458,6 +464,7 @@ class TestFlowKickoffStreaming:
                     LLMStreamChunkEvent(
                         type="llm_stream_chunk",
                         chunk="output!",
+                        call_id="test-call-id",
                     ),
                 )
                 return "done"
@@ -545,6 +552,7 @@ class TestFlowKickoffStreamingAsync:
                     LLMStreamChunkEvent(
                         type="llm_stream_chunk",
                         chunk="Async flow ",
+                        call_id="test-call-id",
                     ),
                 )
                 await asyncio.sleep(0.01)
@@ -553,6 +561,7 @@ class TestFlowKickoffStreamingAsync:
                     LLMStreamChunkEvent(
                         type="llm_stream_chunk",
                         chunk="stream!",
+                        call_id="test-call-id",
                     ),
                 )
                 await asyncio.sleep(0.01)
@@ -686,6 +695,7 @@ class TestStreamingEdgeCases:
                         type="llm_stream_chunk",
                         chunk="Task 1",
                         task_name="First task",
+                        call_id="test-call-id",
                     ),
                 )
                 return mock_output

lib/crewai/tests/utilities/test_events.py

@@ -984,8 +984,8 @@ def test_streaming_fallback_to_non_streaming():
     def mock_call(messages, tools=None, callbacks=None, available_functions=None):
         nonlocal fallback_called
         # Emit a couple of chunks to simulate partial streaming
-        crewai_event_bus.emit(llm, event=LLMStreamChunkEvent(chunk="Test chunk 1", response_id = "Id"))
+        crewai_event_bus.emit(llm, event=LLMStreamChunkEvent(chunk="Test chunk 1", response_id="Id", call_id="test-call-id"))
-        crewai_event_bus.emit(llm, event=LLMStreamChunkEvent(chunk="Test chunk 2", response_id = "Id"))
+        crewai_event_bus.emit(llm, event=LLMStreamChunkEvent(chunk="Test chunk 2", response_id="Id", call_id="test-call-id"))
 
         # Mark that fallback would be called
         fallback_called = True
@@ -1041,7 +1041,7 @@ def test_streaming_empty_response_handling():
     def mock_call(messages, tools=None, callbacks=None, available_functions=None):
         # Emit a few empty chunks
         for _ in range(3):
-            crewai_event_bus.emit(llm, event=LLMStreamChunkEvent(chunk="",response_id="id"))
+            crewai_event_bus.emit(llm, event=LLMStreamChunkEvent(chunk="", response_id="id", call_id="test-call-id"))
 
         # Return the default message for empty responses
         return "I apologize, but I couldn't generate a proper response. Please try again or rephrase your request."
@@ -1280,6 +1280,105 @@ def test_llm_emits_event_with_lite_agent():
     assert set(all_agent_id) == {str(agent.id)}
 
 
+# ----------- CALL_ID CORRELATION TESTS -----------
+
+
+@pytest.mark.vcr()
+def test_llm_call_events_share_call_id():
+    """All events from a single LLM call should share the same call_id."""
+    import uuid
+
+    events = []
+    condition = threading.Condition()
+
+    @crewai_event_bus.on(LLMCallStartedEvent)
+    def on_start(source, event):
+        with condition:
+            events.append(event)
+            condition.notify()
+
+    @crewai_event_bus.on(LLMCallCompletedEvent)
+    def on_complete(source, event):
+        with condition:
+            events.append(event)
+            condition.notify()
+
+    llm = LLM(model="gpt-4o-mini")
+    llm.call("Say hi")
+
+    with condition:
+        success = condition.wait_for(lambda: len(events) >= 2, timeout=10)
+    assert success, "Timeout waiting for LLM events"
+
+    # Behavior: all events from the call share the same call_id
+    assert len(events) == 2
+    assert events[0].call_id == events[1].call_id
+    # call_id should be a valid UUID
+    uuid.UUID(events[0].call_id)
+
+
+@pytest.mark.vcr()
+def test_streaming_chunks_share_call_id_with_call():
+    """Streaming chunks should share call_id with started/completed events."""
+    events = []
+    condition = threading.Condition()
+
+    @crewai_event_bus.on(LLMCallStartedEvent)
+    def on_start(source, event):
+        with condition:
+            events.append(event)
+            condition.notify()
+
+    @crewai_event_bus.on(LLMStreamChunkEvent)
+    def on_chunk(source, event):
+        with condition:
+            events.append(event)
+            condition.notify()
+
+    @crewai_event_bus.on(LLMCallCompletedEvent)
+    def on_complete(source, event):
+        with condition:
+            events.append(event)
+            condition.notify()
+
+    llm = LLM(model="gpt-4o-mini", stream=True)
+    llm.call("Say hi")
+
+    with condition:
+        # Wait for at least started, some chunks, and completed
+        success = condition.wait_for(lambda: len(events) >= 3, timeout=10)
+    assert success, "Timeout waiting for streaming events"
+
+    # Behavior: all events (started, chunks, completed) share the same call_id
+    call_ids = {e.call_id for e in events}
+    assert len(call_ids) == 1
+
+
+@pytest.mark.vcr()
+def test_separate_llm_calls_have_different_call_ids():
+    """Different LLM calls should have different call_ids."""
+    call_ids = []
+    condition = threading.Condition()
+
+    @crewai_event_bus.on(LLMCallStartedEvent)
+    def on_start(source, event):
+        with condition:
+            call_ids.append(event.call_id)
+            condition.notify()
+
+    llm = LLM(model="gpt-4o-mini")
+    llm.call("Say hi")
+    llm.call("Say bye")
+
+    with condition:
+        success = condition.wait_for(lambda: len(call_ids) >= 2, timeout=10)
+    assert success, "Timeout waiting for LLM call events"
+
+    # Behavior: each call has its own call_id
+    assert len(call_ids) == 2
+    assert call_ids[0] != call_ids[1]
+
+
 # ----------- HUMAN FEEDBACK EVENTS -----------