Refine custom embedder configuration support

- Update custom embedder configuration method to handle custom embedding functions
- Modify type hints for embedder configuration
- Remove unused model_name parameter in custom embedder configuration

README.md

@@ -1,18 +1,10 @@
 <div align="center">
 
-![Logo of CrewAI](./docs/crewai_logo.png)
+![Logo of CrewAI, two people rowing on a boat](./docs/crewai_logo.png)
 
 # **CrewAI**
 
-**CrewAI**: Production-grade framework for orchestrating sophisticated AI agent systems. From simple automations to complex real-world applications, CrewAI provides precise control and deep customization. By fostering collaborative intelligence through flexible, production-ready architecture, CrewAI empowers agents to work together seamlessly, tackling complex business challenges with predictable, consistent results.
-
-**CrewAI Enterprise**
-Want to plan, build (+ no code), deploy, monitor and interare your agents: [CrewAI Enterprise](https://www.crewai.com/enterprise). Designed for complex, real-world applications, our enterprise solution offers:
-
-- **Seamless Integrations**
-- **Scalable & Secure Deployment**
-- **Actionable Insights**
-- **24/7 Support**
+🤖 **CrewAI**: Production-grade framework for orchestrating sophisticated AI agent systems. From simple automations to complex real-world applications, CrewAI provides precise control and deep customization. By fostering collaborative intelligence through flexible, production-ready architecture, CrewAI empowers agents to work together seamlessly, tackling complex business challenges with predictable, consistent results.
 
 <h3>
 
@@ -400,7 +392,7 @@ class AdvancedAnalysisFlow(Flow[MarketState]):
             goal="Gather and validate supporting market data",
             backstory="You excel at finding and correlating multiple data sources"
         )
-
+        
         analysis_task = Task(
             description="Analyze {sector} sector data for the past {timeframe}",
             expected_output="Detailed market analysis with confidence score",
@@ -411,7 +403,7 @@ class AdvancedAnalysisFlow(Flow[MarketState]):
             expected_output="Corroborating evidence and potential contradictions",
             agent=researcher
         )
-
+        
         # Demonstrate crew autonomy
         analysis_crew = Crew(
             agents=[analyst, researcher],

docs/concepts/knowledge.mdx

@@ -91,7 +91,7 @@ result = crew.kickoff(inputs={"question": "What city does John live in and how o
 ```
 
 
-Here's another example with the `CrewDoclingSource`. The CrewDoclingSource is actually quite versatile and can handle multiple file formats including MD, PDF, DOCX, HTML, and more. 
+Here's another example with the `CrewDoclingSource`. The CrewDoclingSource is actually quite versatile and can handle multiple file formats including TXT, PDF, DOCX, HTML, and more. 
 
 <Note>
   You need to install `docling` for the following example to work: `uv add docling`
@@ -152,10 +152,10 @@ Here are examples of how to use different types of knowledge sources:
 
 ### Text File Knowledge Source
 ```python
-from crewai.knowledge.source.text_file_knowledge_source import TextFileKnowledgeSource
+from crewai.knowledge.source.crew_docling_source import CrewDoclingSource
 
 # Create a text file knowledge source
-text_source = TextFileKnowledgeSource(
+text_source = CrewDoclingSource(
     file_paths=["document.txt", "another.txt"]
 )
 

docs/concepts/llms.mdx

@@ -463,32 +463,26 @@ Learn how to get the most out of your LLM configuration:
 
   <Accordion title="Google">
     ```python Code
-    # Option 1: Gemini accessed with an API key.
+    # Option 1. Gemini accessed with an API key.
     # https://ai.google.dev/gemini-api/docs/api-key
     GEMINI_API_KEY=<your-api-key>
 
-    # Option 2: Vertex AI IAM credentials for Gemini, Anthropic, and Model Garden.
+    # Option 2. Vertex AI IAM credentials for Gemini, Anthropic, and anything in the Model Garden.
     # https://cloud.google.com/vertex-ai/generative-ai/docs/overview
     ```
 
-    Get credentials:
-    ```python Code
-    import json
-
+    ## GET CREDENTIALS 
     file_path = 'path/to/vertex_ai_service_account.json'
 
     # Load the JSON file
     with open(file_path, 'r') as file:
         vertex_credentials = json.load(file)
 
-    # Convert the credentials to a JSON string
+    # Convert to JSON string
     vertex_credentials_json = json.dumps(vertex_credentials)
-    ```
 
     Example usage:
     ```python Code
-    from crewai import LLM
-
     llm = LLM(
         model="gemini/gemini-1.5-pro-latest",
         temperature=0.7,

docs/concepts/memory.mdx

@@ -58,107 +58,41 @@ my_crew = Crew(
 ### Example: Use Custom Memory Instances e.g FAISS as the VectorDB
 
 ```python Code
-from crewai import Crew, Process
-from crewai.memory import LongTermMemory, ShortTermMemory, EntityMemory
-from crewai.memory.storage import LTMSQLiteStorage, RAGStorage
-from typing import List, Optional
+from crewai import Crew, Agent, Task, Process
 
 # Assemble your crew with memory capabilities
-my_crew: Crew = Crew(
-    agents = [...],
-    tasks = [...],
-    process = Process.sequential,
-    memory = True,
-    # Long-term memory for persistent storage across sessions
-    long_term_memory = LongTermMemory(
+my_crew = Crew(
+    agents=[...],
+    tasks=[...],
+    process="Process.sequential",
+    memory=True,
+    long_term_memory=EnhanceLongTermMemory(
         storage=LTMSQLiteStorage(
-            db_path="/my_crew1/long_term_memory_storage.db"
+            db_path="/my_data_dir/my_crew1/long_term_memory_storage.db"
         )
     ),
-    # Short-term memory for current context using RAG
-    short_term_memory = ShortTermMemory(
-        storage = RAGStorage(
-                embedder_config={
-                    "provider": "openai",
-                    "config": {
-                        "model": 'text-embedding-3-small'
-                    }
-                },
-                type="short_term",
-                path="/my_crew1/"
-            )
+    short_term_memory=EnhanceShortTermMemory(
+        storage=CustomRAGStorage(
+            crew_name="my_crew",
+            storage_type="short_term",
+            data_dir="//my_data_dir",
+            model=embedder["model"],
+            dimension=embedder["dimension"],
         ),
     ),
-    # Entity memory for tracking key information about entities
-    entity_memory = EntityMemory(
-        storage=RAGStorage(
-            embedder_config={
-                "provider": "openai",
-                "config": {
-                    "model": 'text-embedding-3-small'
-                }
-            },
-            type="short_term",
-            path="/my_crew1/"
-        )
+    entity_memory=EnhanceEntityMemory(
+        storage=CustomRAGStorage(
+            crew_name="my_crew",
+            storage_type="entities",
+            data_dir="//my_data_dir",
+            model=embedder["model"],
+            dimension=embedder["dimension"],
+        ),
     ),
     verbose=True,
 )
 ```
 
-## Security Considerations
-
-When configuring memory storage:
-- Use environment variables for storage paths (e.g., `CREWAI_STORAGE_DIR`)
-- Never hardcode sensitive information like database credentials
-- Consider access permissions for storage directories
-- Use relative paths when possible to maintain portability
-
-Example using environment variables:
-```python
-import os
-from crewai import Crew
-from crewai.memory import LongTermMemory
-from crewai.memory.storage import LTMSQLiteStorage
-
-# Configure storage path using environment variable
-storage_path = os.getenv("CREWAI_STORAGE_DIR", "./storage")
-crew = Crew(
-    memory=True,
-    long_term_memory=LongTermMemory(
-        storage=LTMSQLiteStorage(
-            db_path="{storage_path}/memory.db".format(storage_path=storage_path)
-        )
-    )
-)
-```
-
-## Configuration Examples
-
-### Basic Memory Configuration
-```python
-from crewai import Crew
-from crewai.memory import LongTermMemory
-
-# Simple memory configuration
-crew = Crew(memory=True)  # Uses default storage locations
-```
-
-### Custom Storage Configuration
-```python
-from crewai import Crew
-from crewai.memory import LongTermMemory
-from crewai.memory.storage import LTMSQLiteStorage
-
-# Configure custom storage paths
-crew = Crew(
-    memory=True,
-    long_term_memory=LongTermMemory(
-        storage=LTMSQLiteStorage(db_path="./memory.db")
-    )
-)
-```
-
 ## Integrating Mem0 for Enhanced User Memory
 
 [Mem0](https://mem0.ai/) is a self-improving memory layer for LLM applications, enabling personalized AI experiences. 
@@ -282,19 +216,6 @@ my_crew = Crew(
 
 ### Using Google AI embeddings
 
-#### Prerequisites
-Before using Google AI embeddings, ensure you have:
-- Access to the Gemini API
-- The necessary API keys and permissions
-
-You will need to update your *pyproject.toml* dependencies:
-```YAML
-dependencies = [
-    "google-generativeai>=0.8.4", #main version in January/2025 - crewai v.0.100.0 and crewai-tools 0.33.0
-    "crewai[tools]>=0.100.0,<1.0.0"
-]
-```
-
 ```python Code
 from crewai import Crew, Agent, Task, Process
 
@@ -447,38 +368,6 @@ my_crew = Crew(
 )
 ```
 
-### Using Amazon Bedrock embeddings
-
-```python Code
-# Note: Ensure you have installed `boto3` for Bedrock embeddings to work.
-
-import os
-import boto3
-from crewai import Crew, Agent, Task, Process
-
-boto3_session = boto3.Session(
-    region_name=os.environ.get("AWS_REGION_NAME"),
-    aws_access_key_id=os.environ.get("AWS_ACCESS_KEY_ID"),
-    aws_secret_access_key=os.environ.get("AWS_SECRET_ACCESS_KEY")
-)
-
-my_crew = Crew(
-    agents=[...],
-    tasks=[...],
-    process=Process.sequential,
-    memory=True,
-    embedder={
-    "provider": "bedrock",
-        "config":{
-            "session": boto3_session,
-            "model": "amazon.titan-embed-text-v2:0",
-            "vector_dimension": 1024
-        }
-    }
-    verbose=True
-)
-```
-
 ### Adding Custom Embedding Function
 
 ```python Code

docs/concepts/tasks.mdx

@@ -268,7 +268,7 @@ analysis_task = Task(
 
 Task guardrails provide a way to validate and transform task outputs before they
 are passed to the next task. This feature helps ensure data quality and provides
-feedback to agents when their output doesn't meet specific criteria.
+efeedback to agents when their output doesn't meet specific criteria.
 
 ### Using Task Guardrails
 

docs/tools/filewritetool.mdx

@@ -8,9 +8,9 @@ icon: file-pen
 
 ## Description
 
-The `FileWriterTool` is a component of the crewai_tools package, designed to simplify the process of writing content to files with cross-platform compatibility (Windows, Linux, macOS). 
+The `FileWriterTool` is a component of the crewai_tools package, designed to simplify the process of writing content to files. 
 It is particularly useful in scenarios such as generating reports, saving logs, creating configuration files, and more. 
-This tool handles path differences across operating systems, supports UTF-8 encoding, and automatically creates directories if they don't exist, making it easier to organize your output reliably across different platforms.
+This tool supports creating new directories if they don't exist, making it easier to organize your output.
 
 ## Installation
 
@@ -43,8 +43,6 @@ print(result)
 
 ## Conclusion
 
-By integrating the `FileWriterTool` into your crews, the agents can reliably write content to files across different operating systems. 
-This tool is essential for tasks that require saving output data, creating structured file systems, and handling cross-platform file operations. 
-It's particularly recommended for Windows users who may encounter file writing issues with standard Python file operations.
-
-By adhering to the setup and usage guidelines provided, incorporating this tool into projects is straightforward and ensures consistent file writing behavior across all platforms.
+By integrating the `FileWriterTool` into your crews, the agents can execute the process of writing content to files and creating directories. 
+This tool is essential for tasks that require saving output data, creating structured file systems, and more. By adhering to the setup and usage guidelines provided, 
+incorporating this tool into projects is straightforward and efficient.

src/crewai/agent.py

@@ -16,6 +16,7 @@ from crewai.memory.contextual.contextual_memory import ContextualMemory
 from crewai.task import Task
 from crewai.tools import BaseTool
 from crewai.tools.agent_tools.agent_tools import AgentTools
+from crewai.tools.base_tool import Tool
 from crewai.utilities import Converter, Prompts
 from crewai.utilities.constants import TRAINED_AGENTS_DATA_FILE, TRAINING_DATA_FILE
 from crewai.utilities.converter import generate_model_description
@@ -145,7 +146,7 @@ class Agent(BaseAgent):
     def _set_knowledge(self):
         try:
             if self.knowledge_sources:
-                full_pattern = re.compile(r"[^a-zA-Z0-9\-_\r\n]|(\.\.)")
+                full_pattern = re.compile(r'[^a-zA-Z0-9\-_\r\n]|(\.\.)')
                 knowledge_agent_name = f"{re.sub(full_pattern, '_', self.role)}"
                 if isinstance(self.knowledge_sources, list) and all(
                     isinstance(k, BaseKnowledgeSource) for k in self.knowledge_sources

src/crewai/cli/reset_memories_command.py

@@ -3,6 +3,11 @@ import subprocess
 import click
 
 from crewai.cli.utils import get_crew
+from crewai.knowledge.storage.knowledge_storage import KnowledgeStorage
+from crewai.memory.entity.entity_memory import EntityMemory
+from crewai.memory.long_term.long_term_memory import LongTermMemory
+from crewai.memory.short_term.short_term_memory import ShortTermMemory
+from crewai.utilities.task_output_storage_handler import TaskOutputStorageHandler
 
 
 def reset_memories_command(

src/crewai/crew.py

@@ -1,6 +1,7 @@
 import asyncio
 import json
 import re
+import sys
 import uuid
 import warnings
 from concurrent.futures import Future
@@ -380,22 +381,6 @@ class Crew(BaseModel):
 
         return self
 
-    @model_validator(mode="after")
-    def validate_must_have_non_conditional_task(self) -> "Crew":
-        """Ensure that a crew has at least one non-conditional task."""
-        if not self.tasks:
-            return self
-        non_conditional_count = sum(
-            1 for task in self.tasks if not isinstance(task, ConditionalTask)
-        )
-        if non_conditional_count == 0:
-            raise PydanticCustomError(
-                "only_conditional_tasks",
-                "Crew must include at least one non-conditional task",
-                {},
-            )
-        return self
-
     @model_validator(mode="after")
     def validate_first_task(self) -> "Crew":
         """Ensure the first task is not a ConditionalTask."""
@@ -456,7 +441,6 @@ class Crew(BaseModel):
         return self
 
 
-
     @property
     def key(self) -> str:
         source = [agent.key for agent in self.agents] + [
@@ -759,7 +743,6 @@ class Crew(BaseModel):
                     task, task_outputs, futures, task_index, was_replayed
                 )
                 if skipped_task_output:
-                    task_outputs.append(skipped_task_output)
                     continue
 
             if task.async_execution:
@@ -783,7 +766,7 @@ class Crew(BaseModel):
                     context=context,
                     tools=tools_for_task,
                 )
-                task_outputs.append(task_output)
+                task_outputs = [task_output]
                 self._process_task_result(task, task_output)
                 self._store_execution_log(task, task_output, task_index, was_replayed)
 
@@ -804,7 +787,7 @@ class Crew(BaseModel):
             task_outputs = self._process_async_tasks(futures, was_replayed)
             futures.clear()
 
-        previous_output = task_outputs[-1] if task_outputs else None
+        previous_output = task_outputs[task_index - 1] if task_outputs else None
         if previous_output is not None and not task.should_execute(previous_output):
             self._logger.log(
                 "debug",
@@ -926,15 +909,11 @@ class Crew(BaseModel):
             )
 
     def _create_crew_output(self, task_outputs: List[TaskOutput]) -> CrewOutput:
-        if not task_outputs:
-            raise ValueError("No task outputs available to create crew output.")
-            
-        # Filter out empty outputs and get the last valid one as the main output
-        valid_outputs = [t for t in task_outputs if t.raw]
-        if not valid_outputs:
-            raise ValueError("No valid task outputs available to create crew output.")
-        final_task_output = valid_outputs[-1]
-            
+        if len(task_outputs) != 1:
+            raise ValueError(
+                "Something went wrong. Kickoff should return only one task output."
+            )
+        final_task_output = task_outputs[0]
         final_string_output = final_task_output.raw
         self._finish_execution(final_string_output)
         token_usage = self.calculate_usage_metrics()
@@ -943,7 +922,7 @@ class Crew(BaseModel):
             raw=final_task_output.raw,
             pydantic=final_task_output.pydantic,
             json_dict=final_task_output.json_dict,
-            tasks_output=task_outputs,
+            tasks_output=[task.output for task in self.tasks if task.output],
             token_usage=token_usage,
         )
 

src/crewai/knowledge/source/excel_knowledge_source.py

@@ -1,138 +1,28 @@
 from pathlib import Path
-from typing import Dict, Iterator, List, Optional, Union
-from urllib.parse import urlparse
+from typing import Dict, List
 
-from pydantic import Field, field_validator
-
-from crewai.knowledge.source.base_knowledge_source import BaseKnowledgeSource
-from crewai.utilities.constants import KNOWLEDGE_DIRECTORY
-from crewai.utilities.logger import Logger
+from crewai.knowledge.source.base_file_knowledge_source import BaseFileKnowledgeSource
 
 
-class ExcelKnowledgeSource(BaseKnowledgeSource):
+class ExcelKnowledgeSource(BaseFileKnowledgeSource):
     """A knowledge source that stores and queries Excel file content using embeddings."""
 
-    # override content to be a dict of file paths to sheet names to csv content
-
-    _logger: Logger = Logger(verbose=True)
-
-    file_path: Optional[Union[Path, List[Path], str, List[str]]] = Field(
-        default=None,
-        description="[Deprecated] The path to the file. Use file_paths instead.",
-    )
-    file_paths: Optional[Union[Path, List[Path], str, List[str]]] = Field(
-        default_factory=list, description="The path to the file"
-    )
-    chunks: List[str] = Field(default_factory=list)
-    content: Dict[Path, Dict[str, str]] = Field(default_factory=dict)
-    safe_file_paths: List[Path] = Field(default_factory=list)
-
-    @field_validator("file_path", "file_paths", mode="before")
-    def validate_file_path(cls, v, info):
-        """Validate that at least one of file_path or file_paths is provided."""
-        # Single check if both are None, O(1) instead of nested conditions
-        if (
-            v is None
-            and info.data.get(
-                "file_path" if info.field_name == "file_paths" else "file_paths"
-            )
-            is None
-        ):
-            raise ValueError("Either file_path or file_paths must be provided")
-        return v
-
-    def _process_file_paths(self) -> List[Path]:
-        """Convert file_path to a list of Path objects."""
-
-        if hasattr(self, "file_path") and self.file_path is not None:
-            self._logger.log(
-                "warning",
-                "The 'file_path' attribute is deprecated and will be removed in a future version. Please use 'file_paths' instead.",
-                color="yellow",
-            )
-            self.file_paths = self.file_path
-
-        if self.file_paths is None:
-            raise ValueError("Your source must be provided with a file_paths: []")
-
-        # Convert single path to list
-        path_list: List[Union[Path, str]] = (
-            [self.file_paths]
-            if isinstance(self.file_paths, (str, Path))
-            else list(self.file_paths)
-            if isinstance(self.file_paths, list)
-            else []
-        )
-
-        if not path_list:
-            raise ValueError(
-                "file_path/file_paths must be a Path, str, or a list of these types"
-            )
-
-        return [self.convert_to_path(path) for path in path_list]
-
-    def validate_content(self):
-        """Validate the paths."""
-        for path in self.safe_file_paths:
-            if not path.exists():
-                self._logger.log(
-                    "error",
-                    f"File not found: {path}. Try adding sources to the knowledge directory. If it's inside the knowledge directory, use the relative path.",
-                    color="red",
-                )
-                raise FileNotFoundError(f"File not found: {path}")
-            if not path.is_file():
-                self._logger.log(
-                    "error",
-                    f"Path is not a file: {path}",
-                    color="red",
-                )
-
-    def model_post_init(self, _) -> None:
-        if self.file_path:
-            self._logger.log(
-                "warning",
-                "The 'file_path' attribute is deprecated and will be removed in a future version. Please use 'file_paths' instead.",
-                color="yellow",
-            )
-            self.file_paths = self.file_path
-        self.safe_file_paths = self._process_file_paths()
-        self.validate_content()
-        self.content = self._load_content()
-
-    def _load_content(self) -> Dict[Path, Dict[str, str]]:
-        """Load and preprocess Excel file content from multiple sheets.
-
-        Each sheet's content is converted to CSV format and stored.
-
-        Returns:
-            Dict[Path, Dict[str, str]]: A mapping of file paths to their respective sheet contents.
-
-        Raises:
-            ImportError: If required dependencies are missing.
-            FileNotFoundError: If the specified Excel file cannot be opened.
-        """
+    def load_content(self) -> Dict[Path, str]:
+        """Load and preprocess Excel file content."""
         pd = self._import_dependencies()
+
         content_dict = {}
         for file_path in self.safe_file_paths:
             file_path = self.convert_to_path(file_path)
-            with pd.ExcelFile(file_path) as xl:
-                sheet_dict = {
-                    str(sheet_name): str(
-                        pd.read_excel(xl, sheet_name).to_csv(index=False)
-                    )
-                    for sheet_name in xl.sheet_names
-                }
-            content_dict[file_path] = sheet_dict
+            df = pd.read_excel(file_path)
+            content = df.to_csv(index=False)
+            content_dict[file_path] = content
         return content_dict
 
-    def convert_to_path(self, path: Union[Path, str]) -> Path:
-        """Convert a path to a Path object."""
-        return Path(KNOWLEDGE_DIRECTORY + "/" + path) if isinstance(path, str) else path
-
     def _import_dependencies(self):
         """Dynamically import dependencies."""
         try:
+            import openpyxl  # noqa
             import pandas as pd
 
             return pd
@@ -148,14 +38,10 @@ class ExcelKnowledgeSource(BaseKnowledgeSource):
         and save the embeddings.
         """
         # Convert dictionary values to a single string if content is a dictionary
-        # Updated to account for .xlsx workbooks with multiple tabs/sheets
-        content_str = ""
-        for value in self.content.values():
-            if isinstance(value, dict):
-                for sheet_value in value.values():
-                    content_str += str(sheet_value) + "\n"
-            else:
-                content_str += str(value) + "\n"
+        if isinstance(self.content, dict):
+            content_str = "\n".join(str(value) for value in self.content.values())
+        else:
+            content_str = str(self.content)
 
         new_chunks = self._chunk_text(content_str)
         self.chunks.extend(new_chunks)

src/crewai/llm.py

@@ -164,7 +164,6 @@ class LLM:
         self.context_window_size = 0
         self.reasoning_effort = reasoning_effort
         self.additional_params = kwargs
-        self.is_anthropic = self._is_anthropic_model(model)
 
         litellm.drop_params = True
 
@@ -179,62 +178,42 @@ class LLM:
         self.set_callbacks(callbacks)
         self.set_env_callbacks()
 
-    def _is_anthropic_model(self, model: str) -> bool:
-        """Determine if the model is from Anthropic provider.
-        
-        Args:
-            model: The model identifier string.
-            
-        Returns:
-            bool: True if the model is from Anthropic, False otherwise.
-        """
-        ANTHROPIC_PREFIXES = ('anthropic/', 'claude-', 'claude/')
-        return any(prefix in model.lower() for prefix in ANTHROPIC_PREFIXES)
-
     def call(
         self,
         messages: Union[str, List[Dict[str, str]]],
         tools: Optional[List[dict]] = None,
         callbacks: Optional[List[Any]] = None,
         available_functions: Optional[Dict[str, Any]] = None,
-    ) -> Union[str, Any]:
-        """High-level LLM call method.
-        
-        Args:
-            messages: Input messages for the LLM.
-                     Can be a string or list of message dictionaries.
-                     If string, it will be converted to a single user message.
-                     If list, each dict must have 'role' and 'content' keys.
-            tools: Optional list of tool schemas for function calling.
-                  Each tool should define its name, description, and parameters.
-            callbacks: Optional list of callback functions to be executed
-                      during and after the LLM call.
-            available_functions: Optional dict mapping function names to callables
-                               that can be invoked by the LLM.
-        
+    ) -> str:
+        """
+        High-level llm call method that:
+          1) Accepts either a string or a list of messages
+          2) Converts string input to the required message format
+          3) Calls litellm.completion
+          4) Handles function/tool calls if any
+          5) Returns the final text response or tool result
+
+        Parameters:
+        - messages (Union[str, List[Dict[str, str]]]): The input messages for the LLM.
+          - If a string is provided, it will be converted into a message list with a single entry.
+          - If a list of dictionaries is provided, each dictionary should have 'role' and 'content' keys.
+        - tools (Optional[List[dict]]): A list of tool schemas for function calling.
+        - callbacks (Optional[List[Any]]): A list of callback functions to be executed.
+        - available_functions (Optional[Dict[str, Any]]): A dictionary mapping function names to actual Python functions.
+
         Returns:
-            Union[str, Any]: Either a text response from the LLM (str) or
-                           the result of a tool function call (Any).
-        
-        Raises:
-            TypeError: If messages format is invalid
-            ValueError: If response format is not supported
-            LLMContextLengthExceededException: If input exceeds model's context limit
-        
+        - str: The final text response from the LLM or the result of a tool function call.
+
         Examples:
-            # Example 1: Simple string input
-            >>> response = llm.call("Return the name of a random city.")
-            >>> print(response)
-            "Paris"
-            
-            # Example 2: Message list with system and user messages
-            >>> messages = [
-            ...     {"role": "system", "content": "You are a geography expert"},
-            ...     {"role": "user", "content": "What is France's capital?"}
-            ... ]
-            >>> response = llm.call(messages)
-            >>> print(response)
-            "The capital of France is Paris."
+        ---------
+        # Example 1: Using a string input
+        response = llm.call("Return the name of a random city in the world.")
+        print(response)
+
+        # Example 2: Using a list of messages
+        messages = [{"role": "user", "content": "What is the capital of France?"}]
+        response = llm.call(messages)
+        print(response)
         """
         # Validate parameters before proceeding with the call.
         self._validate_call_params()
@@ -254,13 +233,10 @@ class LLM:
                 self.set_callbacks(callbacks)
 
             try:
-                # --- 1) Format messages according to provider requirements
-                formatted_messages = self._format_messages_for_provider(messages)
-
-                # --- 2) Prepare the parameters for the completion call
+                # --- 1) Prepare the parameters for the completion call
                 params = {
                     "model": self.model,
-                    "messages": formatted_messages,
+                    "messages": messages,
                     "timeout": self.timeout,
                     "temperature": self.temperature,
                     "top_p": self.top_p,
@@ -348,38 +324,6 @@ class LLM:
                     logging.error(f"LiteLLM call failed: {str(e)}")
                 raise
 
-    def _format_messages_for_provider(self, messages: List[Dict[str, str]]) -> List[Dict[str, str]]:
-        """Format messages according to provider requirements.
-        
-        Args:
-            messages: List of message dictionaries with 'role' and 'content' keys.
-                     Can be empty or None.
-        
-        Returns:
-            List of formatted messages according to provider requirements.
-            For Anthropic models, ensures first message has 'user' role.
-        
-        Raises:
-            TypeError: If messages is None or contains invalid message format.
-        """
-        if messages is None:
-            raise TypeError("Messages cannot be None")
-            
-        # Validate message format first
-        for msg in messages:
-            if not isinstance(msg, dict) or "role" not in msg or "content" not in msg:
-                raise TypeError("Invalid message format. Each message must be a dict with 'role' and 'content' keys")
-            
-        if not self.is_anthropic:
-            return messages
-            
-        # Anthropic requires messages to start with 'user' role
-        if not messages or messages[0]["role"] == "system":
-            # If first message is system or empty, add a placeholder user message
-            return [{"role": "user", "content": "."}, *messages]
-                
-        return messages
-
     def _get_custom_llm_provider(self) -> str:
         """
         Derives the custom_llm_provider from the model string.

src/crewai/memory/entity/entity_memory.py

@@ -1,4 +1,4 @@
-from typing import Optional
+from typing import Any, Optional
 
 from pydantic import PrivateAttr
 

src/crewai/memory/memory.py

@@ -1,7 +1,9 @@
-from typing import Any, Dict, List, Optional
+from typing import Any, Dict, List, Optional, Union
 
 from pydantic import BaseModel
 
+from crewai.memory.storage.rag_storage import RAGStorage
+
 
 class Memory(BaseModel):
     """

src/crewai/task.py

@@ -674,32 +674,19 @@ class Task(BaseModel):
             return OutputFormat.PYDANTIC
         return OutputFormat.RAW
 
-    def _save_file(self, result: Union[Dict, str, Any]) -> None:
+    def _save_file(self, result: Any) -> None:
         """Save task output to a file.
 
-        Note:
-            For cross-platform file writing, especially on Windows, consider using FileWriterTool
-            from the crewai_tools package:
-                pip install 'crewai[tools]'
-                from crewai_tools import FileWriterTool
-
         Args:
             result: The result to save to the file. Can be a dict or any stringifiable object.
 
         Raises:
             ValueError: If output_file is not set
-            RuntimeError: If there is an error writing to the file. For cross-platform
-                compatibility, especially on Windows, use FileWriterTool from crewai_tools
-                package.
+            RuntimeError: If there is an error writing to the file
         """
         if self.output_file is None:
             raise ValueError("output_file is not set.")
 
-        FILEWRITER_RECOMMENDATION = (
-            "For cross-platform file writing, especially on Windows, "
-            "use FileWriterTool from crewai_tools package."
-        )
-
         try:
             resolved_path = Path(self.output_file).expanduser().resolve()
             directory = resolved_path.parent
@@ -715,12 +702,7 @@ class Task(BaseModel):
                 else:
                     file.write(str(result))
         except (OSError, IOError) as e:
-            raise RuntimeError(
-                "\n".join([
-                    f"Failed to save output file: {e}",
-                    FILEWRITER_RECOMMENDATION
-                ])
-            )
+            raise RuntimeError(f"Failed to save output file: {e}")
         return None
 
     def __repr__(self):

tests/crew_test.py

@@ -49,41 +49,6 @@ writer = Agent(
 )
 
 
-def test_crew_with_only_conditional_tasks_raises_error():
-    """Test that creating a crew with only conditional tasks raises an error."""
-
-    def condition_func(task_output: TaskOutput) -> bool:
-        return True
-
-    conditional1 = ConditionalTask(
-        description="Conditional task 1",
-        expected_output="Output 1",
-        agent=researcher,
-        condition=condition_func,
-    )
-    conditional2 = ConditionalTask(
-        description="Conditional task 2",
-        expected_output="Output 2",
-        agent=researcher,
-        condition=condition_func,
-    )
-    conditional3 = ConditionalTask(
-        description="Conditional task 3",
-        expected_output="Output 3",
-        agent=researcher,
-        condition=condition_func,
-    )
-
-    with pytest.raises(
-        pydantic_core._pydantic_core.ValidationError,
-        match="Crew must include at least one non-conditional task",
-    ):
-        Crew(
-            agents=[researcher],
-            tasks=[conditional1, conditional2, conditional3],
-        )
-
-
 def test_crew_config_conditional_requirement():
     with pytest.raises(ValueError):
         Crew(process=Process.sequential)
@@ -591,12 +556,12 @@ def test_crew_with_delegating_agents_should_not_override_task_tools():
         _, kwargs = mock_execute_sync.call_args
         tools = kwargs["tools"]
 
-        assert any(
-            isinstance(tool, TestTool) for tool in tools
-        ), "TestTool should be present"
-        assert any(
-            "delegate" in tool.name.lower() for tool in tools
-        ), "Delegation tool should be present"
+        assert any(isinstance(tool, TestTool) for tool in tools), (
+            "TestTool should be present"
+        )
+        assert any("delegate" in tool.name.lower() for tool in tools), (
+            "Delegation tool should be present"
+        )
 
 
 @pytest.mark.vcr(filter_headers=["authorization"])
@@ -655,12 +620,12 @@ def test_crew_with_delegating_agents_should_not_override_agent_tools():
         _, kwargs = mock_execute_sync.call_args
         tools = kwargs["tools"]
 
-        assert any(
-            isinstance(tool, TestTool) for tool in new_ceo.tools
-        ), "TestTool should be present"
-        assert any(
-            "delegate" in tool.name.lower() for tool in tools
-        ), "Delegation tool should be present"
+        assert any(isinstance(tool, TestTool) for tool in new_ceo.tools), (
+            "TestTool should be present"
+        )
+        assert any("delegate" in tool.name.lower() for tool in tools), (
+            "Delegation tool should be present"
+        )
 
 
 @pytest.mark.vcr(filter_headers=["authorization"])
@@ -784,17 +749,17 @@ def test_task_tools_override_agent_tools_with_allow_delegation():
         used_tools = kwargs["tools"]
 
         # Confirm AnotherTestTool is present but TestTool is not
-        assert any(
-            isinstance(tool, AnotherTestTool) for tool in used_tools
-        ), "AnotherTestTool should be present"
-        assert not any(
-            isinstance(tool, TestTool) for tool in used_tools
-        ), "TestTool should not be present among used tools"
+        assert any(isinstance(tool, AnotherTestTool) for tool in used_tools), (
+            "AnotherTestTool should be present"
+        )
+        assert not any(isinstance(tool, TestTool) for tool in used_tools), (
+            "TestTool should not be present among used tools"
+        )
 
         # Confirm delegation tool(s) are present
-        assert any(
-            "delegate" in tool.name.lower() for tool in used_tools
-        ), "Delegation tool should be present"
+        assert any("delegate" in tool.name.lower() for tool in used_tools), (
+            "Delegation tool should be present"
+        )
 
     # Finally, make sure the agent's original tools remain unchanged
     assert len(researcher_with_delegation.tools) == 1
@@ -1595,9 +1560,9 @@ def test_code_execution_flag_adds_code_tool_upon_kickoff():
 
         # Verify that exactly one tool was used and it was a CodeInterpreterTool
         assert len(used_tools) == 1, "Should have exactly one tool"
-        assert isinstance(
-            used_tools[0], CodeInterpreterTool
-        ), "Tool should be CodeInterpreterTool"
+        assert isinstance(used_tools[0], CodeInterpreterTool), (
+            "Tool should be CodeInterpreterTool"
+        )
 
 
 @pytest.mark.vcr(filter_headers=["authorization"])
@@ -1954,7 +1919,6 @@ def test_task_callback_on_crew():
 
 def test_task_callback_both_on_task_and_crew():
     from unittest.mock import MagicMock, patch
-
     mock_callback_on_task = MagicMock()
     mock_callback_on_crew = MagicMock()
 
@@ -2096,210 +2060,6 @@ def test_tools_with_custom_caching():
             assert result.raw == "3"
 
 
-@pytest.mark.vcr(filter_headers=["authorization"])
-def test_conditional_task_uses_last_output():
-    """Test that conditional tasks use the last task output for condition evaluation."""
-    task1 = Task(
-        description="First task",
-        expected_output="First output",
-        agent=researcher,
-    )
-
-    def condition_fails(task_output: TaskOutput) -> bool:
-        # This condition will never be met
-        return "never matches" in task_output.raw.lower()
-
-    def condition_succeeds(task_output: TaskOutput) -> bool:
-        # This condition will match first task's output
-        return "first success" in task_output.raw.lower()
-
-    conditional_task1 = ConditionalTask(
-        description="Second task - conditional that fails condition",
-        expected_output="Second output",
-        agent=researcher,
-        condition=condition_fails,
-    )
-
-    conditional_task2 = ConditionalTask(
-        description="Third task - conditional that succeeds using first task output",
-        expected_output="Third output",
-        agent=writer,
-        condition=condition_succeeds,
-    )
-
-    crew = Crew(
-        agents=[researcher, writer],
-        tasks=[task1, conditional_task1, conditional_task2],
-    )
-
-    # Mock outputs for tasks
-    mock_first = TaskOutput(
-        description="First task output",
-        raw="First success output",  # Will be used by third task's condition
-        agent=researcher.role,
-    )
-    mock_third = TaskOutput(
-        description="Third task output",
-        raw="Third task executed",  # Output when condition succeeds using first task output
-        agent=writer.role,
-    )
-
-    # Set up mocks for task execution and conditional logic
-    with patch.object(ConditionalTask, "should_execute") as mock_should_execute:
-        # First conditional fails, second succeeds
-        mock_should_execute.side_effect = [False, True]
-        with patch.object(Task, "execute_sync") as mock_execute:
-            mock_execute.side_effect = [mock_first, mock_third]
-            result = crew.kickoff()
-
-            # Verify execution behavior
-            assert mock_execute.call_count == 2  # Only first and third tasks execute
-            assert mock_should_execute.call_count == 2  # Both conditionals checked
-
-            # Verify outputs collection:
-            # First executed task output, followed by an automatically generated (skipped) output, then the conditional execution
-            assert len(result.tasks_output) == 3
-            assert (
-                result.tasks_output[0].raw == "First success output"
-            )  # First task succeeded
-            assert (
-                result.tasks_output[1].raw == ""
-            )  # Second task skipped (condition failed)
-            assert (
-                result.tasks_output[2].raw == "Third task executed"
-            )  # Third task used first task's output
-
-
-@pytest.mark.vcr(filter_headers=["authorization"])
-def test_conditional_tasks_result_collection():
-    """Test that task outputs are properly collected based on execution status."""
-    task1 = Task(
-        description="Normal task that always executes",
-        expected_output="First output",
-        agent=researcher,
-    )
-
-    def condition_never_met(task_output: TaskOutput) -> bool:
-        return "never matches" in task_output.raw.lower()
-
-    def condition_always_met(task_output: TaskOutput) -> bool:
-        return "success" in task_output.raw.lower()
-
-    task2 = ConditionalTask(
-        description="Conditional task that never executes",
-        expected_output="Second output",
-        agent=researcher,
-        condition=condition_never_met,
-    )
-
-    task3 = ConditionalTask(
-        description="Conditional task that always executes",
-        expected_output="Third output",
-        agent=writer,
-        condition=condition_always_met,
-    )
-
-    crew = Crew(
-        agents=[researcher, writer],
-        tasks=[task1, task2, task3],
-    )
-
-    # Mock outputs for different execution paths
-    mock_success = TaskOutput(
-        description="Success output",
-        raw="Success output",  # Triggers third task's condition
-        agent=researcher.role,
-    )
-    mock_conditional = TaskOutput(
-        description="Conditional output",
-        raw="Conditional task executed",
-        agent=writer.role,
-    )
-
-    # Set up mocks for task execution and conditional logic
-    with patch.object(ConditionalTask, "should_execute") as mock_should_execute:
-        # First conditional fails, second succeeds
-        mock_should_execute.side_effect = [False, True]
-        with patch.object(Task, "execute_sync") as mock_execute:
-            mock_execute.side_effect = [mock_success, mock_conditional]
-            result = crew.kickoff()
-
-            # Verify execution behavior
-            assert mock_execute.call_count == 2  # Only first and third tasks execute
-            assert mock_should_execute.call_count == 2  # Both conditionals checked
-
-            # Verify task output collection:
-            # There should be three outputs: normal task, skipped conditional task (empty output),
-            # and the conditional task that executed.
-            assert len(result.tasks_output) == 3
-            assert (
-                result.tasks_output[0].raw == "Success output"
-            )  # Normal task executed
-            assert result.tasks_output[1].raw == ""  # Second task skipped
-            assert (
-                result.tasks_output[2].raw == "Conditional task executed"
-            )  # Third task executed
-
-            # Verify task output collection
-            assert len(result.tasks_output) == 3
-            assert (
-                result.tasks_output[0].raw == "Success output"
-            )  # Normal task executed
-            assert result.tasks_output[1].raw == ""  # Second task skipped
-            assert (
-                result.tasks_output[2].raw == "Conditional task executed"
-            )  # Third task executed
-
-
-@pytest.mark.vcr(filter_headers=["authorization"])
-def test_multiple_conditional_tasks():
-    """Test that having multiple conditional tasks in sequence works correctly."""
-    task1 = Task(
-        description="Initial research task",
-        expected_output="Research output",
-        agent=researcher,
-    )
-
-    def condition1(task_output: TaskOutput) -> bool:
-        return "success" in task_output.raw.lower()
-
-    def condition2(task_output: TaskOutput) -> bool:
-        return "proceed" in task_output.raw.lower()
-
-    task2 = ConditionalTask(
-        description="First conditional task",
-        expected_output="Conditional output 1",
-        agent=writer,
-        condition=condition1,
-    )
-
-    task3 = ConditionalTask(
-        description="Second conditional task",
-        expected_output="Conditional output 2",
-        agent=writer,
-        condition=condition2,
-    )
-
-    crew = Crew(
-        agents=[researcher, writer],
-        tasks=[task1, task2, task3],
-    )
-
-    # Mock different task outputs to test conditional logic
-    mock_success = TaskOutput(
-        description="Mock success",
-        raw="Success and proceed output",
-        agent=researcher.role,
-    )
-
-    # Set up mocks for task execution
-    with patch.object(Task, "execute_sync", return_value=mock_success) as mock_execute:
-        result = crew.kickoff()
-        # Verify all tasks were executed (no IndexError)
-        assert mock_execute.call_count == 3
-        assert len(result.tasks_output) == 3
-
-
 @pytest.mark.vcr(filter_headers=["authorization"])
 def test_using_contextual_memory():
     from unittest.mock import patch
@@ -3418,9 +3178,9 @@ def test_fetch_inputs():
     expected_placeholders = {"role_detail", "topic", "field"}
     actual_placeholders = crew.fetch_inputs()
 
-    assert (
-        actual_placeholders == expected_placeholders
-    ), f"Expected {expected_placeholders}, but got {actual_placeholders}"
+    assert actual_placeholders == expected_placeholders, (
+        f"Expected {expected_placeholders}, but got {actual_placeholders}"
+    )
 
 
 def test_task_tools_preserve_code_execution_tools():
@@ -3493,20 +3253,20 @@ def test_task_tools_preserve_code_execution_tools():
         used_tools = kwargs["tools"]
 
         # Verify all expected tools are present
-        assert any(
-            isinstance(tool, TestTool) for tool in used_tools
-        ), "Task's TestTool should be present"
-        assert any(
-            isinstance(tool, CodeInterpreterTool) for tool in used_tools
-        ), "CodeInterpreterTool should be present"
-        assert any(
-            "delegate" in tool.name.lower() for tool in used_tools
-        ), "Delegation tool should be present"
+        assert any(isinstance(tool, TestTool) for tool in used_tools), (
+            "Task's TestTool should be present"
+        )
+        assert any(isinstance(tool, CodeInterpreterTool) for tool in used_tools), (
+            "CodeInterpreterTool should be present"
+        )
+        assert any("delegate" in tool.name.lower() for tool in used_tools), (
+            "Delegation tool should be present"
+        )
 
         # Verify the total number of tools (TestTool + CodeInterpreter + 2 delegation tools)
-        assert (
-            len(used_tools) == 4
-        ), "Should have TestTool, CodeInterpreter, and 2 delegation tools"
+        assert len(used_tools) == 4, (
+            "Should have TestTool, CodeInterpreter, and 2 delegation tools"
+        )
 
 
 @pytest.mark.vcr(filter_headers=["authorization"])
@@ -3550,9 +3310,9 @@ def test_multimodal_flag_adds_multimodal_tools():
         used_tools = kwargs["tools"]
 
         # Check that the multimodal tool was added
-        assert any(
-            isinstance(tool, AddImageTool) for tool in used_tools
-        ), "AddImageTool should be present when agent is multimodal"
+        assert any(isinstance(tool, AddImageTool) for tool in used_tools), (
+            "AddImageTool should be present when agent is multimodal"
+        )
 
         # Verify we have exactly one tool (just the AddImageTool)
         assert len(used_tools) == 1, "Should only have the AddImageTool"
@@ -3778,9 +3538,9 @@ def test_crew_guardrail_feedback_in_context():
     assert len(execution_contexts) > 1, "Task should have been executed multiple times"
 
     # Verify that the second execution included the guardrail feedback
-    assert (
-        "Output must contain the keyword 'IMPORTANT'" in execution_contexts[1]
-    ), "Guardrail feedback should be included in retry context"
+    assert "Output must contain the keyword 'IMPORTANT'" in execution_contexts[1], (
+        "Guardrail feedback should be included in retry context"
+    )
 
     # Verify final output meets guardrail requirements
     assert "IMPORTANT" in result.raw, "Final output should contain required keyword"

tests/llm_test.py

@@ -286,79 +286,6 @@ def test_o3_mini_reasoning_effort_medium():
 
 
 @pytest.mark.vcr(filter_headers=["authorization"])
-@pytest.fixture
-def anthropic_llm():
-    """Fixture providing an Anthropic LLM instance."""
-    return LLM(model="anthropic/claude-3-sonnet")
-
-@pytest.fixture
-def system_message():
-    """Fixture providing a system message."""
-    return {"role": "system", "content": "test"}
-
-@pytest.fixture
-def user_message():
-    """Fixture providing a user message."""
-    return {"role": "user", "content": "test"}
-
-def test_anthropic_message_formatting_edge_cases(anthropic_llm):
-    """Test edge cases for Anthropic message formatting."""
-    # Test None messages
-    with pytest.raises(TypeError, match="Messages cannot be None"):
-        anthropic_llm._format_messages_for_provider(None)
-        
-    # Test empty message list
-    formatted = anthropic_llm._format_messages_for_provider([])
-    assert len(formatted) == 1
-    assert formatted[0]["role"] == "user"
-    assert formatted[0]["content"] == "."
-    
-    # Test invalid message format
-    with pytest.raises(TypeError, match="Invalid message format"):
-        anthropic_llm._format_messages_for_provider([{"invalid": "message"}])
-
-def test_anthropic_model_detection():
-    """Test Anthropic model detection with various formats."""
-    models = [
-        ("anthropic/claude-3", True),
-        ("claude-instant", True),
-        ("claude/v1", True),
-        ("gpt-4", False),
-        ("", False),
-        ("anthropomorphic", False),  # Should not match partial words
-    ]
-    
-    for model, expected in models:
-        llm = LLM(model=model)
-        assert llm.is_anthropic == expected, f"Failed for model: {model}"
-
-def test_anthropic_message_formatting(anthropic_llm, system_message, user_message):
-    """Test Anthropic message formatting with fixtures."""
-    # Test when first message is system
-    formatted = anthropic_llm._format_messages_for_provider([system_message])
-    assert len(formatted) == 2
-    assert formatted[0]["role"] == "user"
-    assert formatted[0]["content"] == "."
-    assert formatted[1] == system_message
-
-    # Test when first message is already user
-    formatted = anthropic_llm._format_messages_for_provider([user_message])
-    assert len(formatted) == 1
-    assert formatted[0] == user_message
-
-    # Test with empty message list
-    formatted = anthropic_llm._format_messages_for_provider([])
-    assert len(formatted) == 1
-    assert formatted[0]["role"] == "user"
-    assert formatted[0]["content"] == "."
-
-    # Test with non-Anthropic model (should not modify messages)
-    non_anthropic_llm = LLM(model="gpt-4")
-    formatted = non_anthropic_llm._format_messages_for_provider([system_message])
-    assert len(formatted) == 1
-    assert formatted[0] == system_message
-
-
 def test_deepseek_r1_with_open_router():
     if not os.getenv("OPEN_ROUTER_API_KEY"):
         pytest.skip("OPEN_ROUTER_API_KEY not set; skipping test.")