Implement comprehensive streaming support for CrewAI

- Add streaming events: CrewStreamChunkEvent, TaskStreamChunkEvent, AgentStreamChunkEvent
- Extend Crew.kickoff() with stream parameter and callback support
- Propagate streaming through task and agent execution chains
- Integrate with existing LLM streaming infrastructure
- Add comprehensive tests and examples
- Maintain backward compatibility

Fixes #2950

Co-Authored-By: João <joao@crewai.com>

README_STREAMING.md

@@ -0,0 +1,151 @@
+# CrewAI Streaming Support
+
+This document describes the streaming functionality added to CrewAI to support real-time output during crew execution.
+
+## Overview
+
+The streaming feature allows users to receive real-time updates during crew execution, similar to how autogen and langgraph provide streaming capabilities. This is particularly useful for multi-agent scenarios where you want to see the progress of each agent and task as they execute.
+
+## Usage
+
+### Basic Streaming
+
+```python
+from crewai import Agent, Task, Crew
+from crewai.llm import LLM
+
+def stream_callback(chunk, agent_role, task_description, step_type):
+    """Callback function to handle streaming chunks."""
+    print(f"[{agent_role}] {step_type}: {chunk}", end="", flush=True)
+
+llm = LLM(model="gpt-4o-mini", stream=True)
+
+agent = Agent(
+    role="Content Writer",
+    goal="Write engaging content",
+    backstory="You are an experienced content writer.",
+    llm=llm
+)
+
+task = Task(
+    description="Write a short story about AI",
+    expected_output="A creative short story",
+    agent=agent
+)
+
+crew = Crew(
+    agents=[agent],
+    tasks=[task]
+)
+
+# Enable streaming with callback
+result = crew.kickoff(
+    stream=True,
+    stream_callback=stream_callback
+)
+```
+
+### Multi-Agent Streaming
+
+```python
+def stream_callback(chunk, agent_role, task_description, step_type):
+    """Enhanced callback for multi-agent scenarios."""
+    timestamp = datetime.now().strftime("%H:%M:%S")
+    print(f"[{timestamp}] {agent_role} ({step_type}): {chunk}", end="", flush=True)
+
+researcher = Agent(
+    role="Research Analyst",
+    goal="Research topics thoroughly",
+    backstory="You are an experienced researcher.",
+    llm=llm
+)
+
+writer = Agent(
+    role="Content Writer", 
+    goal="Write based on research",
+    backstory="You create compelling content.",
+    llm=llm
+)
+
+research_task = Task(
+    description="Research AI trends",
+    expected_output="Research summary",
+    agent=researcher
+)
+
+writing_task = Task(
+    description="Write blog post about AI trends",
+    expected_output="Blog post",
+    agent=writer,
+    context=[research_task]
+)
+
+crew = Crew(
+    agents=[researcher, writer],
+    tasks=[research_task, writing_task]
+)
+
+result = crew.kickoff(
+    stream=True,
+    stream_callback=stream_callback
+)
+```
+
+## API Reference
+
+### Crew.kickoff()
+
+```python
+def kickoff(
+    self,
+    inputs: Optional[Dict[str, Any]] = None,
+    stream: bool = False,
+    stream_callback: Optional[Callable[[str, str, str, str], None]] = None,
+) -> CrewOutput:
+```
+
+**Parameters:**
+- `inputs`: Dictionary of inputs for the crew
+- `stream`: Whether to enable streaming output (default: False)
+- `stream_callback`: Callback function for streaming chunks
+
+**Stream Callback Signature:**
+```python
+def stream_callback(chunk: str, agent_role: str, task_description: str, step_type: str) -> None:
+```
+
+**Callback Parameters:**
+- `chunk`: The streaming text chunk
+- `agent_role`: Role of the agent producing the chunk
+- `task_description`: Description of the current task
+- `step_type`: Type of step ("agent_thinking", "final_answer", "llm_response", etc.)
+
+## Events
+
+The streaming system emits several types of events:
+
+### CrewStreamChunkEvent
+Emitted for crew-level streaming chunks with context about the agent and task.
+
+### TaskStreamChunkEvent  
+Emitted for task-level streaming chunks.
+
+### AgentStreamChunkEvent
+Emitted for agent-level streaming chunks.
+
+## Integration with Existing LLM Streaming
+
+The crew streaming builds on top of the existing LLM streaming infrastructure. When you enable streaming at the crew level, it automatically aggregates and contextualizes the LLM-level streaming chunks.
+
+## Best Practices
+
+1. **Enable LLM Streaming**: Make sure your LLM has `stream=True` for optimal experience
+2. **Handle Empty Chunks**: Your callback should handle empty or whitespace-only chunks gracefully
+3. **Performance**: Streaming adds minimal overhead but consider disabling for batch processing
+4. **Error Handling**: Implement proper error handling in your stream callback
+
+## Examples
+
+See the `examples/` directory for complete working examples:
+- `streaming_example.py`: Basic single-agent streaming
+- `streaming_multi_agent_example.py`: Multi-agent streaming with context

docs/streaming.md

@@ -0,0 +1,68 @@
+# Streaming Support in CrewAI
+
+CrewAI now supports real-time streaming output during crew execution, allowing you to see the progress of agents and tasks as they work.
+
+## Basic Usage
+
+```python
+from crewai import Agent, Task, Crew
+from crewai.llm import LLM
+
+def stream_callback(chunk, agent_role, task_description, step_type):
+    print(f"[{agent_role}] {step_type}: {chunk}", end="", flush=True)
+
+llm = LLM(model="gpt-4o-mini", stream=True)
+
+agent = Agent(
+    role="Writer",
+    goal="Write content",
+    backstory="You are a skilled writer.",
+    llm=llm
+)
+
+task = Task(
+    description="Write a short story",
+    expected_output="A creative story",
+    agent=agent
+)
+
+crew = Crew(agents=[agent], tasks=[task])
+
+result = crew.kickoff(
+    stream=True,
+    stream_callback=stream_callback
+)
+```
+
+## Multi-Agent Streaming
+
+```python
+def enhanced_callback(chunk, agent_role, task_description, step_type):
+    print(f"[{agent_role}] {task_description[:20]}... - {step_type}: {chunk}")
+
+researcher = Agent(role="Researcher", ...)
+writer = Agent(role="Writer", ...)
+
+research_task = Task(description="Research topic", agent=researcher)
+write_task = Task(description="Write article", agent=writer, context=[research_task])
+
+crew = Crew(agents=[researcher, writer], tasks=[research_task, write_task])
+result = crew.kickoff(stream=True, stream_callback=enhanced_callback)
+```
+
+## Stream Callback Parameters
+
+- `chunk`: The streaming text chunk
+- `agent_role`: Role of the agent producing the chunk  
+- `task_description`: Description of the current task
+- `step_type`: Type of step ("agent_thinking", "final_answer", "llm_response")
+
+## Events
+
+The streaming system emits `CrewStreamChunkEvent`, `TaskStreamChunkEvent`, and `AgentStreamChunkEvent` that can be handled using the event bus.
+
+## Requirements
+
+- Enable streaming on your LLM: `LLM(model="...", stream=True)`
+- Use the `stream=True` parameter in `crew.kickoff()`
+- Provide a callback function to handle streaming chunks

examples/streaming_example.py

@@ -0,0 +1,36 @@
+from crewai import Agent, Task, Crew
+from crewai.llm import LLM
+
+def stream_callback(chunk, agent_role, task_description, step_type):
+    """Callback function to handle streaming chunks."""
+    print(f"[{agent_role}] {step_type}: {chunk}", end="", flush=True)
+
+llm = LLM(model="gpt-4o-mini", stream=True)
+
+agent = Agent(
+    role="Content Writer",
+    goal="Write engaging content",
+    backstory="You are an experienced content writer who creates compelling narratives.",
+    llm=llm,
+    verbose=False
+)
+
+task = Task(
+    description="Write a short story about a robot learning to paint",
+    expected_output="A creative short story of 2-3 paragraphs",
+    agent=agent
+)
+
+crew = Crew(
+    agents=[agent],
+    tasks=[task],
+    verbose=False
+)
+
+print("Starting crew execution with streaming...")
+result = crew.kickoff(
+    stream=True,
+    stream_callback=stream_callback
+)
+
+print(f"\n\nFinal result:\n{result}")

examples/streaming_multi_agent_example.py

@@ -0,0 +1,51 @@
+from crewai import Agent, Task, Crew
+from crewai.llm import LLM
+
+def stream_callback(chunk, agent_role, task_description, step_type):
+    """Callback function to handle streaming chunks from multiple agents."""
+    print(f"[{agent_role}] {step_type}: {chunk}", end="", flush=True)
+
+llm = LLM(model="gpt-4o-mini", stream=True)
+
+researcher = Agent(
+    role="Research Analyst",
+    goal="Research and analyze topics thoroughly",
+    backstory="You are an experienced research analyst who excels at gathering and analyzing information.",
+    llm=llm,
+    verbose=False
+)
+
+writer = Agent(
+    role="Content Writer", 
+    goal="Write engaging content based on research",
+    backstory="You are a skilled content writer who creates compelling narratives from research data.",
+    llm=llm,
+    verbose=False
+)
+
+research_task = Task(
+    description="Research the latest trends in artificial intelligence and machine learning",
+    expected_output="A comprehensive research summary of AI/ML trends",
+    agent=researcher
+)
+
+writing_task = Task(
+    description="Write an engaging blog post about AI trends based on the research",
+    expected_output="A well-written blog post about AI trends",
+    agent=writer,
+    context=[research_task]
+)
+
+crew = Crew(
+    agents=[researcher, writer],
+    tasks=[research_task, writing_task],
+    verbose=False
+)
+
+print("Starting multi-agent crew execution with streaming...")
+result = crew.kickoff(
+    stream=True,
+    stream_callback=stream_callback
+)
+
+print(f"\n\nFinal result:\n{result}")

src/crewai/agent.py

@@ -1,6 +1,6 @@
 import shutil
 import subprocess
-from typing import Any, Dict, List, Literal, Optional, Sequence, Type, Union
+from typing import Any, Callable, Dict, List, Literal, Optional, Sequence, Type, Union
 
 from pydantic import Field, InstanceOf, PrivateAttr, model_validator
 
@@ -225,6 +225,8 @@ class Agent(BaseAgent):
         task: Task,
         context: Optional[str] = None,
         tools: Optional[List[BaseTool]] = None,
+        stream: bool = False,
+        stream_callback: Optional[Callable[[str, str, str, str], None]] = None,
     ) -> str:
         """Execute a task with the agent.
 
@@ -232,6 +234,8 @@ class Agent(BaseAgent):
             task: Task to execute.
             context: Context to execute the task in.
             tools: Tools to use for the task.
+            stream: Whether to enable streaming output.
+            stream_callback: Callback function for streaming chunks.
 
         Returns:
             Output of the agent
@@ -364,6 +368,10 @@ class Agent(BaseAgent):
         tools = tools or self.tools or []
         self.create_agent_executor(tools=tools, task=task)
         
+        if stream and stream_callback:
+            self.agent_executor._stream_callback = stream_callback
+            self.agent_executor._task_description = task.description
+
         if self.crew and self.crew._train:
             task_prompt = self._training_handler(task_prompt=task_prompt)
         else:
@@ -429,7 +437,7 @@ class Agent(BaseAgent):
                     ),
                 )
                 raise e
-            result = self.execute_task(task, context, tools)
+            result = self.execute_task(task, context, tools, stream, stream_callback)
 
         if self.max_rpm and self._rpm_controller:
             self._rpm_controller.stop_rpm_counter()

src/crewai/agents/crew_agent_executor.py

@@ -80,6 +80,8 @@ class CrewAgentExecutor(CrewAgentExecutorMixin):
         self.messages: List[Dict[str, str]] = []
         self.iterations = 0
         self.log_error_after = 3
+        self._stream_callback = None
+        self._task_description = None
         self.tool_name_to_tool_map: Dict[str, Union[CrewStructuredTool, BaseTool]] = {
             tool.name: tool for tool in self.tools
         }
@@ -158,6 +160,23 @@ class CrewAgentExecutor(CrewAgentExecutorMixin):
                 )
                 formatted_answer = process_llm_response(answer, self.use_stop_words)
                 
+                if hasattr(self, '_stream_callback') and self._stream_callback:
+                    if hasattr(formatted_answer, 'text'):
+                        step_type = "agent_thinking" if hasattr(formatted_answer, 'tool') else "final_answer"
+                        self._stream_callback(
+                            formatted_answer.text,
+                            self.agent.role if self.agent else "unknown",
+                            getattr(self, '_task_description', "unknown"),
+                            step_type
+                        )
+                    elif isinstance(formatted_answer, str):
+                        self._stream_callback(
+                            formatted_answer,
+                            self.agent.role if self.agent else "unknown",
+                            getattr(self, '_task_description', "unknown"),
+                            "final_answer"
+                        )
+
                 if isinstance(formatted_answer, AgentAction):
                     # Extract agent fingerprint if available
                     fingerprint_context = {}

src/crewai/crew.py

@@ -615,13 +615,30 @@ class Crew(FlowTrackable, BaseModel):
     def kickoff(
         self,
         inputs: Optional[Dict[str, Any]] = None,
+        stream: bool = False,
+        stream_callback: Optional[Callable[[str, str, str, str], None]] = None,
     ) -> CrewOutput:
+        """
+        Starts the crew to work on its assigned tasks.
+
+        Args:
+            inputs (dict): Inputs to be used by the crew.
+            stream (bool): Whether to enable streaming output.
+            stream_callback (callable): Callback function for streaming chunks.
+                                      Signature: (chunk, agent_role, task_description, step_type)
+
+        Returns:
+            CrewOutput: The output of the crew.
+        """
         try:
             for before_callback in self.before_kickoff_callbacks:
                 if inputs is None:
                     inputs = {}
                 inputs = before_callback(inputs)
 
+            self._stream_enabled = stream
+            self._stream_callback = stream_callback
+
             crewai_event_bus.emit(
                 self,
                 CrewKickoffStartedEvent(crew_name=self.name or "crew", inputs=inputs),
@@ -865,6 +882,8 @@ class Crew(FlowTrackable, BaseModel):
                     agent=agent_to_use,
                     context=context,
                     tools=cast(List[BaseTool], tools_for_task),
+                    stream=getattr(self, '_stream_enabled', False),
+                    stream_callback=getattr(self, '_stream_callback', None),
                 )
                 futures.append((task, future, task_index))
             else:
@@ -877,6 +896,8 @@ class Crew(FlowTrackable, BaseModel):
                     agent=agent_to_use,
                     context=context,
                     tools=cast(List[BaseTool], tools_for_task),
+                    stream=getattr(self, '_stream_enabled', False),
+                    stream_callback=getattr(self, '_stream_callback', None),
                 )
                 task_outputs.append(task_output)
                 self._process_task_result(task, task_output)

src/crewai/task.py

@@ -346,9 +346,11 @@ class Task(BaseModel):
         agent: Optional[BaseAgent] = None,
         context: Optional[str] = None,
         tools: Optional[List[BaseTool]] = None,
+        stream: bool = False,
+        stream_callback: Optional[Callable[[str, str, str, str], None]] = None,
     ) -> TaskOutput:
         """Execute the task synchronously."""
-        return self._execute_core(agent, context, tools)
+        return self._execute_core(agent, context, tools, stream, stream_callback)
 
     @property
     def key(self) -> str:
@@ -369,13 +371,15 @@ class Task(BaseModel):
         agent: BaseAgent | None = None,
         context: Optional[str] = None,
         tools: Optional[List[BaseTool]] = None,
+        stream: bool = False,
+        stream_callback: Optional[Callable[[str, str, str, str], None]] = None,
     ) -> Future[TaskOutput]:
         """Execute the task asynchronously."""
         future: Future[TaskOutput] = Future()
         threading.Thread(
             daemon=True,
             target=self._execute_task_async,
-            args=(agent, context, tools, future),
+            args=(agent, context, tools, future, stream, stream_callback),
         ).start()
         return future
 
@@ -385,9 +389,11 @@ class Task(BaseModel):
         context: Optional[str],
         tools: Optional[List[Any]],
         future: Future[TaskOutput],
+        stream: bool = False,
+        stream_callback: Optional[Callable[[str, str, str, str], None]] = None,
     ) -> None:
         """Execute the task asynchronously with context handling."""
-        result = self._execute_core(agent, context, tools)
+        result = self._execute_core(agent, context, tools, stream, stream_callback)
         future.set_result(result)
 
     def _execute_core(
@@ -395,6 +401,8 @@ class Task(BaseModel):
         agent: Optional[BaseAgent],
         context: Optional[str],
         tools: Optional[List[Any]],
+        stream: bool = False,
+        stream_callback: Optional[Callable[[str, str, str, str], None]] = None,
     ) -> TaskOutput:
         """Run the core execution logic of the task."""
         try:
@@ -416,6 +424,8 @@ class Task(BaseModel):
                 task=self,
                 context=context,
                 tools=tools,
+                stream=stream,
+                stream_callback=stream_callback,
             )
 
             pydantic_output, json_output = self._export_output(result)
@@ -449,7 +459,7 @@ class Task(BaseModel):
                         content=f"Guardrail blocked, retrying, due to: {guardrail_result.error}\n",
                         color="yellow",
                     )
-                    return self._execute_core(agent, context, tools)
+                    return self._execute_core(agent, context, tools, stream, stream_callback)
 
                 if guardrail_result.result is None:
                     raise Exception(

src/crewai/utilities/events/crew_events.py

@@ -109,3 +109,32 @@ class CrewTestResultEvent(CrewBaseEvent):
     execution_duration: float
     model: str
     type: str = "crew_test_result"
+
+
+class CrewStreamChunkEvent(CrewBaseEvent):
+    """Event emitted when a streaming chunk is received during crew execution"""
+    
+    type: str = "crew_stream_chunk"
+    chunk: str
+    agent_role: Optional[str] = None
+    task_description: Optional[str] = None
+    step_type: str
+
+
+class TaskStreamChunkEvent(BaseEvent):
+    """Event emitted when a streaming chunk is received during task execution"""
+    
+    type: str = "task_stream_chunk"
+    chunk: str
+    task_description: str
+    agent_role: str
+    step_type: str
+
+
+class AgentStreamChunkEvent(BaseEvent):
+    """Event emitted when a streaming chunk is received during agent execution"""
+    
+    type: str = "agent_stream_chunk"
+    chunk: str
+    agent_role: str
+    step_type: str

src/crewai/utilities/events/event_listener.py

@@ -358,6 +358,34 @@ class EventListener(BaseEventListener):
             print(content, end="", flush=True)
             self.next_chunk = self.text_stream.tell()
             
+            from crewai.utilities.events.crew_events import CrewStreamChunkEvent
+            from crewai.utilities.events.crewai_event_bus import crewai_event_bus
+            
+            agent_role = "unknown"
+            task_description = "unknown"
+            
+            if hasattr(source, 'agent') and source.agent:
+                agent_role = source.agent.role
+            elif hasattr(source, 'role'):
+                agent_role = source.role
+                
+            if hasattr(source, 'task') and source.task:
+                task_description = source.task.description
+            elif hasattr(source, '_task_description'):
+                task_description = source._task_description
+                
+            crewai_event_bus.emit(
+                source,
+                CrewStreamChunkEvent(
+                    chunk=event.chunk,
+                    agent_role=agent_role,
+                    task_description=task_description,
+                    step_type="llm_response",
+                    crew=getattr(source, 'crew', None),
+                    crew_name=getattr(source, 'crew', {}).get('__class__', {}).get('__name__', None) if hasattr(source, 'crew') and source.crew else None
+                )
+            )
+
         @crewai_event_bus.on(CrewTestStartedEvent)
         def on_crew_test_started(source, event: CrewTestStartedEvent):
             cloned_crew = source.copy()

tests/test_streaming.py

@@ -0,0 +1,261 @@
+import pytest
+from unittest.mock import Mock, patch
+from crewai import Agent, Task, Crew
+from crewai.llm import LLM
+from crewai.utilities.events.crew_events import CrewStreamChunkEvent, TaskStreamChunkEvent, AgentStreamChunkEvent
+from crewai.utilities.events.crewai_event_bus import crewai_event_bus
+
+
+@pytest.fixture
+def mock_llm():
+    return Mock()
+
+
+@pytest.fixture
+def agent(mock_llm):
+    return Agent(
+        role="Test Agent",
+        goal="Test goal",
+        backstory="Test backstory",
+        llm=mock_llm,
+        verbose=False
+    )
+
+
+@pytest.fixture
+def task(agent):
+    return Task(
+        description="Test task",
+        expected_output="Test output",
+        agent=agent
+    )
+
+
+@pytest.fixture
+def crew(agent, task):
+    return Crew(
+        agents=[agent],
+        tasks=[task],
+        verbose=False
+    )
+
+
+def test_crew_streaming_enabled():
+    """Test that crew streaming can be enabled."""
+    received_chunks = []
+    
+    def stream_callback(chunk, agent_role, task_desc, step_type):
+        received_chunks.append({
+            'chunk': chunk,
+            'agent_role': agent_role,
+            'task_desc': task_desc,
+            'step_type': step_type
+        })
+    
+    with patch('crewai.llm.LLM') as mock_llm_class:
+        mock_llm = Mock()
+        mock_llm.call.return_value = "Test response"
+        mock_llm_class.return_value = mock_llm
+        
+        agent = Agent(
+            role="Test Agent",
+            goal="Test goal", 
+            backstory="Test backstory",
+            llm=mock_llm,
+            verbose=False
+        )
+        
+        task = Task(
+            description="Test task",
+            expected_output="Test output",
+            agent=agent
+        )
+        
+        crew = Crew(
+            agents=[agent],
+            tasks=[task],
+            verbose=False
+        )
+        
+        with patch.object(crew, '_execute_tasks') as mock_execute:
+            mock_execute.return_value = Mock()
+            
+            crew.kickoff(stream=True, stream_callback=stream_callback)
+            
+            assert hasattr(crew, '_stream_enabled')
+            assert crew._stream_enabled is True
+            assert hasattr(crew, '_stream_callback')
+            assert crew._stream_callback == stream_callback
+
+
+def test_crew_streaming_disabled_by_default():
+    """Test that crew streaming is disabled by default."""
+    with patch('crewai.llm.LLM') as mock_llm_class:
+        mock_llm = Mock()
+        mock_llm.call.return_value = "Test response"
+        mock_llm_class.return_value = mock_llm
+        
+        agent = Agent(
+            role="Test Agent",
+            goal="Test goal",
+            backstory="Test backstory", 
+            llm=mock_llm,
+            verbose=False
+        )
+        
+        task = Task(
+            description="Test task",
+            expected_output="Test output",
+            agent=agent
+        )
+        
+        crew = Crew(
+            agents=[agent],
+            tasks=[task],
+            verbose=False
+        )
+        
+        with patch.object(crew, '_execute_tasks') as mock_execute:
+            mock_execute.return_value = Mock()
+            
+            crew.kickoff()
+            
+            assert getattr(crew, '_stream_enabled', False) is False
+            assert getattr(crew, '_stream_callback', None) is None
+
+
+def test_crew_stream_chunk_event():
+    """Test CrewStreamChunkEvent creation and properties."""
+    event = CrewStreamChunkEvent(
+        chunk="test chunk",
+        agent_role="Test Agent",
+        task_description="Test task",
+        step_type="agent_thinking",
+        crew=None,
+        crew_name="TestCrew"
+    )
+    
+    assert event.type == "crew_stream_chunk"
+    assert event.chunk == "test chunk"
+    assert event.agent_role == "Test Agent"
+    assert event.task_description == "Test task"
+    assert event.step_type == "agent_thinking"
+
+
+def test_task_stream_chunk_event():
+    """Test TaskStreamChunkEvent creation and properties."""
+    event = TaskStreamChunkEvent(
+        chunk="test chunk",
+        task_description="Test task",
+        agent_role="Test Agent",
+        step_type="task_execution"
+    )
+    
+    assert event.type == "task_stream_chunk"
+    assert event.chunk == "test chunk"
+    assert event.task_description == "Test task"
+    assert event.agent_role == "Test Agent"
+    assert event.step_type == "task_execution"
+
+
+def test_agent_stream_chunk_event():
+    """Test AgentStreamChunkEvent creation and properties."""
+    event = AgentStreamChunkEvent(
+        chunk="test chunk",
+        agent_role="Test Agent",
+        step_type="agent_thinking"
+    )
+    
+    assert event.type == "agent_stream_chunk"
+    assert event.chunk == "test chunk"
+    assert event.agent_role == "Test Agent"
+    assert event.step_type == "agent_thinking"
+
+
+def test_streaming_integration_with_llm():
+    """Test that streaming integrates with existing LLM streaming."""
+    received_callback_chunks = []
+    
+    def stream_callback(chunk, agent_role, task_desc, step_type):
+        received_callback_chunks.append({
+            'chunk': chunk,
+            'agent_role': agent_role,
+            'task_desc': task_desc,
+            'step_type': step_type
+        })
+    
+    with patch('crewai.llm.LLM') as mock_llm_class:
+        mock_llm = Mock()
+        mock_llm.call.return_value = "Here's a joke: Why did the robot cross the road? To get to the other side!"
+        mock_llm_class.return_value = mock_llm
+        
+        agent = Agent(
+            role="Test Agent",
+            goal="Test goal",
+            backstory="Test backstory",
+            llm=mock_llm,
+            verbose=False
+        )
+        
+        task = Task(
+            description="Tell me a short joke",
+            expected_output="A short joke",
+            agent=agent
+        )
+        
+        crew = Crew(
+            agents=[agent],
+            tasks=[task],
+            verbose=False
+        )
+        
+        with patch.object(agent, 'agent_executor') as mock_executor:
+            mock_executor._stream_callback = None
+            mock_executor._task_description = None
+            
+            result = crew.kickoff(stream=True, stream_callback=stream_callback)
+            
+            assert hasattr(agent.agent_executor, '_stream_callback')
+            assert hasattr(agent.agent_executor, '_task_description')
+
+
+def test_streaming_parameters_propagation():
+    """Test that streaming parameters are properly propagated through the execution chain."""
+    with patch('crewai.llm.LLM') as mock_llm_class:
+        mock_llm = Mock()
+        mock_llm.call.return_value = "Test response"
+        mock_llm_class.return_value = mock_llm
+        
+        agent = Agent(
+            role="Test Agent",
+            goal="Test goal",
+            backstory="Test backstory",
+            llm=mock_llm,
+            verbose=False
+        )
+        
+        task = Task(
+            description="Test task",
+            expected_output="Test output",
+            agent=agent
+        )
+        
+        crew = Crew(
+            agents=[agent],
+            tasks=[task],
+            verbose=False
+        )
+        
+        stream_callback = Mock()
+        
+        with patch.object(task, 'execute_sync') as mock_execute_sync:
+            mock_execute_sync.return_value = Mock()
+            
+            crew.kickoff(stream=True, stream_callback=stream_callback)
+            
+            mock_execute_sync.assert_called_once()
+            call_args = mock_execute_sync.call_args
+            assert 'stream' in call_args.kwargs
+            assert call_args.kwargs['stream'] is True
+            assert 'stream_callback' in call_args.kwargs
+            assert call_args.kwargs['stream_callback'] == stream_callback

tests/test_streaming_comprehensive.py

@@ -0,0 +1,271 @@
+import pytest
+from unittest.mock import Mock, patch, MagicMock
+from crewai import Agent, Task, Crew
+from crewai.utilities.events.crew_events import CrewStreamChunkEvent
+from crewai.utilities.events.llm_events import LLMStreamChunkEvent
+from crewai.utilities.events.crewai_event_bus import crewai_event_bus
+
+
+def test_streaming_callback_called():
+    """Test that streaming callback is called during execution."""
+    callback_calls = []
+    
+    def stream_callback(chunk, agent_role, task_desc, step_type):
+        callback_calls.append({
+            'chunk': chunk,
+            'agent_role': agent_role,
+            'task_desc': task_desc,
+            'step_type': step_type
+        })
+    
+    with patch('crewai.llm.LLM') as mock_llm_class:
+        mock_llm = Mock()
+        mock_llm.call.return_value = "Test response"
+        mock_llm_class.return_value = mock_llm
+        
+        agent = Agent(
+            role="Test Agent",
+            goal="Test goal",
+            backstory="Test backstory",
+            llm=mock_llm,
+            verbose=False
+        )
+        
+        task = Task(
+            description="Test task",
+            expected_output="Test output",
+            agent=agent
+        )
+        
+        crew = Crew(
+            agents=[agent],
+            tasks=[task],
+            verbose=False
+        )
+        
+        with patch.object(agent, 'agent_executor') as mock_executor:
+            mock_executor._stream_callback = None
+            mock_executor._task_description = None
+            
+            crew.kickoff(stream=True, stream_callback=stream_callback)
+            
+            assert hasattr(agent.agent_executor, '_stream_callback')
+            assert agent.agent_executor._stream_callback == stream_callback
+            assert hasattr(agent.agent_executor, '_task_description')
+            assert agent.agent_executor._task_description == "Test task"
+
+
+def test_crew_stream_chunk_event_creation():
+    """Test CrewStreamChunkEvent can be created with all required fields."""
+    event = CrewStreamChunkEvent(
+        chunk="test chunk",
+        agent_role="Test Agent",
+        task_description="Test task",
+        step_type="agent_thinking",
+        crew=None,
+        crew_name="TestCrew"
+    )
+    
+    assert event.type == "crew_stream_chunk"
+    assert event.chunk == "test chunk"
+    assert event.agent_role == "Test Agent"
+    assert event.task_description == "Test task"
+    assert event.step_type == "agent_thinking"
+
+
+def test_streaming_disabled_by_default():
+    """Test that streaming is disabled by default."""
+    with patch('crewai.llm.LLM') as mock_llm_class:
+        mock_llm = Mock()
+        mock_llm.call.return_value = "Test response"
+        mock_llm_class.return_value = mock_llm
+        
+        agent = Agent(
+            role="Test Agent",
+            goal="Test goal",
+            backstory="Test backstory",
+            llm=mock_llm,
+            verbose=False
+        )
+        
+        task = Task(
+            description="Test task",
+            expected_output="Test output",
+            agent=agent
+        )
+        
+        crew = Crew(
+            agents=[agent],
+            tasks=[task],
+            verbose=False
+        )
+        
+        crew.kickoff()
+        
+        assert getattr(crew, '_stream_enabled', False) is False
+        assert getattr(crew, '_stream_callback', None) is None
+
+
+def test_streaming_parameters_propagation():
+    """Test that streaming parameters are propagated through execution chain."""
+    stream_callback = Mock()
+    
+    with patch('crewai.llm.LLM') as mock_llm_class:
+        mock_llm = Mock()
+        mock_llm.call.return_value = "Test response"
+        mock_llm_class.return_value = mock_llm
+        
+        agent = Agent(
+            role="Test Agent",
+            goal="Test goal",
+            backstory="Test backstory",
+            llm=mock_llm,
+            verbose=False
+        )
+        
+        task = Task(
+            description="Test task",
+            expected_output="Test output",
+            agent=agent
+        )
+        
+        crew = Crew(
+            agents=[agent],
+            tasks=[task],
+            verbose=False
+        )
+        
+        with patch.object(task, 'execute_sync') as mock_execute_sync:
+            mock_execute_sync.return_value = Mock()
+            
+            crew.kickoff(stream=True, stream_callback=stream_callback)
+            
+            mock_execute_sync.assert_called_once()
+            call_args = mock_execute_sync.call_args
+            assert 'stream' in call_args.kwargs
+            assert call_args.kwargs['stream'] is True
+            assert 'stream_callback' in call_args.kwargs
+            assert call_args.kwargs['stream_callback'] == stream_callback
+
+
+def test_async_task_streaming():
+    """Test that streaming works with async tasks."""
+    stream_callback = Mock()
+    
+    with patch('crewai.llm.LLM') as mock_llm_class:
+        mock_llm = Mock()
+        mock_llm.call.return_value = "Test response"
+        mock_llm_class.return_value = mock_llm
+        
+        agent = Agent(
+            role="Test Agent",
+            goal="Test goal",
+            backstory="Test backstory",
+            llm=mock_llm,
+            verbose=False
+        )
+        
+        task = Task(
+            description="Test task",
+            expected_output="Test output",
+            agent=agent,
+            async_execution=True
+        )
+        
+        crew = Crew(
+            agents=[agent],
+            tasks=[task],
+            verbose=False
+        )
+        
+        with patch.object(task, 'execute_async') as mock_execute_async:
+            mock_future = Mock()
+            mock_execute_async.return_value = mock_future
+            
+            crew.kickoff(stream=True, stream_callback=stream_callback)
+            
+            mock_execute_async.assert_called_once()
+            call_args = mock_execute_async.call_args
+            assert 'stream' in call_args.kwargs
+            assert call_args.kwargs['stream'] is True
+            assert 'stream_callback' in call_args.kwargs
+            assert call_args.kwargs['stream_callback'] == stream_callback
+
+
+def test_llm_stream_chunk_to_crew_stream_chunk():
+    """Test that LLMStreamChunkEvent triggers CrewStreamChunkEvent."""
+    received_crew_chunks = []
+    
+    with crewai_event_bus.scoped_handlers():
+        @crewai_event_bus.on(CrewStreamChunkEvent)
+        def handle_crew_stream_chunk(source, event):
+            received_crew_chunks.append(event)
+        
+        mock_source = Mock()
+        mock_source.agent = Mock()
+        mock_source.agent.role = "Test Agent"
+        mock_source._task_description = "Test task"
+        
+        llm_event = LLMStreamChunkEvent(chunk="test chunk")
+        
+        from crewai.utilities.events.event_listener import event_listener
+        event_listener.on_llm_stream_chunk(mock_source, llm_event)
+        
+        assert len(received_crew_chunks) == 1
+        crew_event = received_crew_chunks[0]
+        assert crew_event.type == "crew_stream_chunk"
+        assert crew_event.chunk == "test chunk"
+        assert crew_event.agent_role == "Test Agent"
+        assert crew_event.task_description == "Test task"
+        assert crew_event.step_type == "llm_response"
+
+
+def test_multiple_agents_streaming():
+    """Test streaming with multiple agents."""
+    stream_callback = Mock()
+    
+    with patch('crewai.llm.LLM') as mock_llm_class:
+        mock_llm = Mock()
+        mock_llm.call.return_value = "Agent response"
+        mock_llm_class.return_value = mock_llm
+        
+        agent1 = Agent(
+            role="Agent 1",
+            goal="Goal 1",
+            backstory="Backstory 1",
+            llm=mock_llm,
+            verbose=False
+        )
+        
+        agent2 = Agent(
+            role="Agent 2",
+            goal="Goal 2",
+            backstory="Backstory 2",
+            llm=mock_llm,
+            verbose=False
+        )
+        
+        task1 = Task(
+            description="Task 1",
+            expected_output="Output 1",
+            agent=agent1
+        )
+        
+        task2 = Task(
+            description="Task 2",
+            expected_output="Output 2",
+            agent=agent2
+        )
+        
+        crew = Crew(
+            agents=[agent1, agent2],
+            tasks=[task1, task2],
+            verbose=False
+        )
+        
+        result = crew.kickoff(stream=True, stream_callback=stream_callback)
+        
+        assert hasattr(crew, '_stream_enabled')
+        assert crew._stream_enabled is True
+        assert hasattr(crew, '_stream_callback')
+        assert crew._stream_callback == stream_callback

tests/test_streaming_integration.py

@@ -0,0 +1,215 @@
+import pytest
+from unittest.mock import Mock, patch
+from crewai import Agent, Task, Crew
+from crewai.utilities.events.crew_events import CrewStreamChunkEvent
+from crewai.utilities.events.llm_events import LLMStreamChunkEvent
+from crewai.utilities.events.crewai_event_bus import crewai_event_bus
+
+
+def test_streaming_callback_integration():
+    """Test that streaming callback is properly integrated through the execution chain."""
+    received_chunks = []
+    
+    def stream_callback(chunk, agent_role, task_desc, step_type):
+        received_chunks.append({
+            'chunk': chunk,
+            'agent_role': agent_role,
+            'task_desc': task_desc,
+            'step_type': step_type
+        })
+    
+    with patch('crewai.llm.LLM') as mock_llm_class:
+        mock_llm = Mock()
+        mock_llm.call.return_value = "Test response"
+        mock_llm_class.return_value = mock_llm
+        
+        agent = Agent(
+            role="Test Agent",
+            goal="Test goal",
+            backstory="Test backstory",
+            llm=mock_llm,
+            verbose=False
+        )
+        
+        task = Task(
+            description="Test task",
+            expected_output="Test output",
+            agent=agent
+        )
+        
+        crew = Crew(
+            agents=[agent],
+            tasks=[task],
+            verbose=False
+        )
+        
+        with patch.object(agent, 'agent_executor') as mock_executor:
+            mock_executor._stream_callback = None
+            mock_executor._task_description = None
+            
+            crew.kickoff(stream=True, stream_callback=stream_callback)
+            
+            assert hasattr(agent.agent_executor, '_stream_callback')
+            assert hasattr(agent.agent_executor, '_task_description')
+
+
+def test_crew_stream_chunk_event_emission():
+    """Test that CrewStreamChunkEvent is emitted when LLMStreamChunkEvent occurs."""
+    received_crew_chunks = []
+    
+    with crewai_event_bus.scoped_handlers():
+        @crewai_event_bus.on(CrewStreamChunkEvent)
+        def handle_crew_stream_chunk(source, event):
+            received_crew_chunks.append(event)
+        
+        mock_source = Mock()
+        mock_source.agent = Mock()
+        mock_source.agent.role = "Test Agent"
+        mock_source._task_description = "Test task"
+        
+        llm_event = LLMStreamChunkEvent(chunk="test chunk")
+        
+        from crewai.utilities.events.event_listener import event_listener
+        event_listener.on_llm_stream_chunk(mock_source, llm_event)
+        
+        assert len(received_crew_chunks) == 1
+        crew_event = received_crew_chunks[0]
+        assert crew_event.type == "crew_stream_chunk"
+        assert crew_event.chunk == "test chunk"
+        assert crew_event.agent_role == "Test Agent"
+        assert crew_event.task_description == "Test task"
+        assert crew_event.step_type == "llm_response"
+
+
+def test_streaming_with_multiple_agents():
+    """Test streaming works correctly with multiple agents."""
+    received_chunks = []
+    
+    def stream_callback(chunk, agent_role, task_desc, step_type):
+        received_chunks.append({
+            'chunk': chunk,
+            'agent_role': agent_role,
+            'task_desc': task_desc,
+            'step_type': step_type
+        })
+    
+    with patch('crewai.llm.LLM') as mock_llm_class:
+        mock_llm = Mock()
+        mock_llm.call.return_value = "Agent response"
+        mock_llm_class.return_value = mock_llm
+        
+        agent1 = Agent(
+            role="Agent 1",
+            goal="Goal 1",
+            backstory="Backstory 1",
+            llm=mock_llm,
+            verbose=False
+        )
+        
+        agent2 = Agent(
+            role="Agent 2", 
+            goal="Goal 2",
+            backstory="Backstory 2",
+            llm=mock_llm,
+            verbose=False
+        )
+        
+        task1 = Task(
+            description="Task 1",
+            expected_output="Output 1",
+            agent=agent1
+        )
+        
+        task2 = Task(
+            description="Task 2",
+            expected_output="Output 2",
+            agent=agent2
+        )
+        
+        crew = Crew(
+            agents=[agent1, agent2],
+            tasks=[task1, task2],
+            verbose=False
+        )
+        
+        result = crew.kickoff(stream=True, stream_callback=stream_callback)
+        
+        assert hasattr(crew, '_stream_enabled')
+        assert crew._stream_enabled is True
+        assert hasattr(crew, '_stream_callback')
+        assert crew._stream_callback == stream_callback
+
+
+def test_streaming_disabled_by_default():
+    """Test that streaming is disabled by default."""
+    with patch('crewai.llm.LLM') as mock_llm_class:
+        mock_llm = Mock()
+        mock_llm.call.return_value = "Test response"
+        mock_llm_class.return_value = mock_llm
+        
+        agent = Agent(
+            role="Test Agent",
+            goal="Test goal",
+            backstory="Test backstory",
+            llm=mock_llm,
+            verbose=False
+        )
+        
+        task = Task(
+            description="Test task",
+            expected_output="Test output",
+            agent=agent
+        )
+        
+        crew = Crew(
+            agents=[agent],
+            tasks=[task],
+            verbose=False
+        )
+        
+        crew.kickoff()
+        
+        assert getattr(crew, '_stream_enabled', False) is False
+        assert getattr(crew, '_stream_callback', None) is None
+
+
+def test_streaming_parameters_propagation():
+    """Test that streaming parameters are properly propagated through execution chain."""
+    stream_callback = Mock()
+    
+    with patch('crewai.llm.LLM') as mock_llm_class:
+        mock_llm = Mock()
+        mock_llm.call.return_value = "Test response"
+        mock_llm_class.return_value = mock_llm
+        
+        agent = Agent(
+            role="Test Agent",
+            goal="Test goal",
+            backstory="Test backstory",
+            llm=mock_llm,
+            verbose=False
+        )
+        
+        task = Task(
+            description="Test task",
+            expected_output="Test output",
+            agent=agent
+        )
+        
+        crew = Crew(
+            agents=[agent],
+            tasks=[task],
+            verbose=False
+        )
+        
+        with patch.object(task, 'execute_sync') as mock_execute_sync:
+            mock_execute_sync.return_value = Mock()
+            
+            crew.kickoff(stream=True, stream_callback=stream_callback)
+            
+            mock_execute_sync.assert_called_once()
+            call_args = mock_execute_sync.call_args
+            assert 'stream' in call_args.kwargs
+            assert call_args.kwargs['stream'] is True
+            assert 'stream_callback' in call_args.kwargs
+            assert call_args.kwargs['stream_callback'] == stream_callback

tests/utilities/test_events.py

@@ -779,3 +779,54 @@ def test_streaming_empty_response_handling():
         finally:
             # Restore the original method
             llm.call = original_call
+
+
+@pytest.mark.vcr(filter_headers=["authorization"])
+def test_crew_streaming_events():
+    """Test that crew streaming events are emitted correctly."""
+    from crewai.utilities.events.crew_events import CrewStreamChunkEvent
+    received_crew_chunks = []
+
+    with crewai_event_bus.scoped_handlers():
+
+        @crewai_event_bus.on(CrewStreamChunkEvent)
+        def handle_crew_stream_chunk(source, event):
+            received_crew_chunks.append(event)
+
+        # Create an LLM with streaming enabled
+        llm = LLM(model="gpt-4o-mini", stream=True)
+
+        # Create agent and task
+        from crewai import Agent, Task, Crew
+        
+        agent = Agent(
+            role="Test Agent",
+            goal="Test goal",
+            backstory="Test backstory",
+            llm=llm,
+            verbose=False
+        )
+        
+        task = Task(
+            description="Tell me a short joke",
+            expected_output="A short joke",
+            agent=agent
+        )
+        
+        crew = Crew(
+            agents=[agent],
+            tasks=[task],
+            verbose=False
+        )
+
+        # Execute with streaming enabled
+        result = crew.kickoff(stream=True)
+
+        # Verify that we received crew stream chunks
+        assert len(received_crew_chunks) > 0
+        
+        for chunk_event in received_crew_chunks:
+            assert chunk_event.type == "crew_stream_chunk"
+            assert chunk_event.agent_role == "Test Agent"
+            assert chunk_event.step_type == "llm_response"
+            assert isinstance(chunk_event.chunk, str)