agno 1.8.1__py3-none-any.whl → 2.0.0__py3-none-any.whl

agno/memory/manager.py

@@ -1,218 +1,1073 @@
-from typing import Any, Dict, List, Optional, cast
+from copy import deepcopy
+from dataclasses import dataclass
+from datetime import datetime
+from os import getenv
+from textwrap import dedent
+from typing import Any, Callable, Dict, List, Literal, Optional, Type, Union
 
-from pydantic import BaseModel, ConfigDict
+from pydantic import BaseModel, Field
 
-from agno.memory.db import MemoryDb
-from agno.memory.memory import Memory
-from agno.memory.row import MemoryRow
+from agno.db.base import BaseDb
+from agno.db.schemas import UserMemory
 from agno.models.base import Model
 from agno.models.message import Message
 from agno.tools.function import Function
-from agno.utils.log import log_debug, logger
+from agno.utils.log import log_debug, log_error, log_warning, set_log_level_to_debug, set_log_level_to_info
+from agno.utils.prompts import get_json_output_prompt
+from agno.utils.string import parse_response_model_str
 
 
-class MemoryManager(BaseModel):
+class MemorySearchResponse(BaseModel):
+    """Model for Memory Search Response."""
+
+    memory_ids: List[str] = Field(
+        ..., description="The IDs of the memories that are most semantically similar to the query."
+    )
+
+
+@dataclass
+class MemoryManager:
+    """Memory Manager"""
+
+    # Model used for memory management
     model: Optional[Model] = None
-    user_id: Optional[str] = None
-    limit: Optional[int] = None
-    # Provide the system prompt for the manager as a string
-    system_prompt: Optional[str] = None
-    # Memory Database
-    db: Optional[MemoryDb] = None
-
-    # Do not set the input message here, it will be set by the run method
-    input_message: Optional[str] = None
-    _tools_for_model: Optional[List[Dict]] = None
-    _functions_for_model: Optional[Dict[str, Function]] = None
-
-    model_config = ConfigDict(arbitrary_types_allowed=True)
-
-    def update_model(self) -> None:
-        # Use the default Model (OpenAIChat) if no model is provided
+
+    # Provide the system message for the manager as a string. If not provided, a default prompt will be used.
+    system_message: Optional[str] = None
+    # Provide the memory capture instructions for the manager as a string. If not provided, a default prompt will be used.
+    memory_capture_instructions: Optional[str] = None
+    # Additional instructions for the manager
+    additional_instructions: Optional[str] = None
+
+    # Whether memories were created in the last run
+    memories_updated: bool = False
+
+    # ----- db tools ---------
+    # Whether to delete memories
+    delete_memories: bool = True
+    # Whether to clear memories
+    clear_memories: bool = True
+    # Whether to update memories
+    update_memories: bool = True
+    # whether to add memories
+    add_memories: bool = True
+
+    # The database to store memories
+    db: Optional[BaseDb] = None
+
+    debug_mode: bool = False
+
+    def __init__(
+        self,
+        model: Optional[Model] = None,
+        system_message: Optional[str] = None,
+        memory_capture_instructions: Optional[str] = None,
+        additional_instructions: Optional[str] = None,
+        db: Optional[BaseDb] = None,
+        delete_memories: bool = True,
+        update_memories: bool = True,
+        add_memories: bool = True,
+        clear_memories: bool = True,
+        debug_mode: bool = False,
+    ):
+        self.model = model
+        if self.model is not None and isinstance(self.model, str):
+            raise ValueError("Model must be a Model object, not a string")
+        self.system_message = system_message
+        self.memory_capture_instructions = memory_capture_instructions
+        self.additional_instructions = additional_instructions
+        self.db = db
+        self.delete_memories = delete_memories
+        self.update_memories = update_memories
+        self.add_memories = add_memories
+        self.clear_memories = clear_memories
+        self.debug_mode = debug_mode
+        self._tools_for_model: Optional[List[Dict[str, Any]]] = None
+        self._functions_for_model: Optional[Dict[str, Function]] = None
+
+    def get_model(self) -> Model:
         if self.model is None:
             try:
                 from agno.models.openai import OpenAIChat
             except ModuleNotFoundError as e:
-                logger.exception(e)
-                logger.error(
+                log_error(e)
+                log_error(
                     "Agno uses `openai` as the default model provider. Please provide a `model` or install `openai`."
                 )
                 exit(1)
             self.model = OpenAIChat(id="gpt-4o")
+        return self.model
 
-    def determine_tools_for_model(self) -> None:
-        if self._tools_for_model is None:
-            self._tools_for_model = []
-        if self._functions_for_model is None:
-            self._functions_for_model = {}
-
-        for tool in [
-            self.add_memory,
-            self.update_memory,
-            self.delete_memory,
-            self.clear_memory,
-        ]:
-            try:
-                function_name = tool.__name__
-                if function_name not in self._functions_for_model:
-                    func = Function.from_callable(tool)  # type: ignore
-                    self._functions_for_model[func.name] = func
-                    self._tools_for_model.append({"type": "function", "function": func.to_dict()})
-                    log_debug(f"Added function {func.name}")
-            except Exception as e:
-                logger.warning(f"Could not add function {tool}: {e}")
+    def read_from_db(self, user_id: Optional[str] = None):
+        if self.db:
+            # If no user_id is provided, read all memories
+            if user_id is None:
+                all_memories: List[UserMemory] = self.db.get_user_memories()  # type: ignore
+            else:
+                all_memories = self.db.get_user_memories(user_id=user_id)  # type: ignore
 
-    def get_existing_memories(self) -> Optional[List[MemoryRow]]:
-        if self.db is None:
+            memories: Dict[str, List[UserMemory]] = {}
+            for memory in all_memories:
+                if memory.user_id is not None and memory.memory_id is not None:
+                    memories.setdefault(memory.user_id, []).append(memory)
+
+            return memories
+        return None
+
+    def set_log_level(self):
+        if self.debug_mode or getenv("AGNO_DEBUG", "false").lower() == "true":
+            self.debug_mode = True
+            set_log_level_to_debug()
+        else:
+            set_log_level_to_info()
+
+    def initialize(self, user_id: Optional[str] = None):
+        self.set_log_level()
+
+    # -*- Public Functions
+    def get_user_memories(self, user_id: Optional[str] = None) -> Optional[List[UserMemory]]:
+        """Get the user memories for a given user id"""
+        if self.db:
+            if user_id is None:
+                user_id = "default"
+            # Refresh from the Db
+            memories = self.read_from_db(user_id=user_id)
+            if memories is None:
+                return []
+            return memories.get(user_id, [])
+        else:
+            log_warning("Memory Db not provided.")
+            return []
+
+    def get_user_memory(self, memory_id: str, user_id: Optional[str] = None) -> Optional[UserMemory]:
+        """Get the user memory for a given user id"""
+        if self.db:
+            if user_id is None:
+                user_id = "default"
+            # Refresh from the DB
+            memories = self.read_from_db(user_id=user_id)
+            if memories is None:
+                return None
+            memories_for_user = memories.get(user_id, [])
+            for memory in memories_for_user:
+                if memory.memory_id == memory_id:
+                    return memory
+            return None
+        else:
+            log_warning("Memory Db not provided.")
             return None
 
-        return self.db.read_memories(user_id=self.user_id, limit=self.limit)
+    def add_user_memory(
+        self,
+        memory: UserMemory,
+        user_id: Optional[str] = None,
+    ) -> Optional[str]:
+        """Add a user memory for a given user id
+        Args:
+            memory (UserMemory): The memory to add
+            user_id (Optional[str]): The user id to add the memory to. If not provided, the memory is added to the "default" user.
+        Returns:
+            str: The id of the memory
+        """
+        if self.db:
+            if memory.memory_id is None:
+                from uuid import uuid4
 
-    def add_memory(self, memory: str) -> str:
-        """Use this function to add a memory to the database.
+                memory_id = memory.memory_id or str(uuid4())
+                memory.memory_id = memory_id
+
+            if user_id is None:
+                user_id = "default"
+            memory.user_id = user_id
+
+            if not memory.updated_at:
+                memory.updated_at = datetime.now()
+
+            self._upsert_db_memory(memory=memory)
+            return memory.memory_id
+
+        else:
+            log_warning("Memory Db not provided.")
+            return None
+
+    def replace_user_memory(
+        self,
+        memory_id: str,
+        memory: UserMemory,
+        user_id: Optional[str] = None,
+    ) -> Optional[str]:
+        """Replace a user memory for a given user id
         Args:
-            memory (str): The memory to be stored.
+            memory_id (str): The id of the memory to replace
+            memory (UserMemory): The memory to add
+            user_id (Optional[str]): The user id to add the memory to. If not provided, the memory is added to the "default" user.
         Returns:
-            str: A message indicating if the memory was added successfully or not.
+            str: The id of the memory
+        """
+        if self.db:
+            if user_id is None:
+                user_id = "default"
+
+            if not memory.updated_at:
+                memory.updated_at = datetime.now()
+
+            memory.memory_id = memory_id
+            memory.user_id = user_id
+
+            self._upsert_db_memory(memory=memory)
+
+            return memory.memory_id
+        else:
+            log_warning("Memory Db not provided.")
+            return None
+
+    def clear(self) -> None:
+        """Clears the memory."""
+        if self.db:
+            self.db.clear_memories()
+
+    def delete_user_memory(
+        self,
+        memory_id: str,
+        user_id: Optional[str] = None,
+    ) -> None:
+        """Delete a user memory for a given user id
+        Args:
+            memory_id (str): The id of the memory to delete
+            user_id (Optional[str]): The user id to delete the memory from. If not provided, the memory is deleted from the "default" user.
         """
+        if self.db:
+            self._delete_db_memory(memory_id=memory_id)
+        else:
+            log_warning("Memory DB not provided.")
+            return None
+
+    # -*- Agent Functions
+    def create_user_memories(
+        self,
+        message: Optional[str] = None,
+        messages: Optional[List[Message]] = None,
+        agent_id: Optional[str] = None,
+        team_id: Optional[str] = None,
+        user_id: Optional[str] = None,
+    ) -> str:
+        """Creates memories from multiple messages and adds them to the memory db."""
+        self.set_log_level()
+
+        if self.db is None:
+            log_warning("MemoryDb not provided.")
+            return "Please provide a db to store memories"
+
+        if not messages and not message:
+            raise ValueError("You must provide either a message or a list of messages")
+
+        if message:
+            messages = [Message(role="user", content=message)]
+
+        if not messages or not isinstance(messages, list):
+            raise ValueError("Invalid messages list")
+
+        if user_id is None:
+            user_id = "default"
+
+        memories = self.read_from_db(user_id=user_id)
+        if memories is None:
+            memories = {}
+
+        existing_memories = memories.get(user_id, [])  # type: ignore
+        existing_memories = [{"memory_id": memory.memory_id, "memory": memory.memory} for memory in existing_memories]
+        response = self.create_or_update_memories(  # type: ignore
+            messages=messages,
+            existing_memories=existing_memories,
+            user_id=user_id,
+            agent_id=agent_id,
+            team_id=team_id,
+            db=self.db,
+            update_memories=self.update_memories,
+            add_memories=self.add_memories,
+        )
+
+        # We refresh from the DB
+        self.read_from_db(user_id=user_id)
+        return response
+
+    async def acreate_user_memories(
+        self,
+        message: Optional[str] = None,
+        messages: Optional[List[Message]] = None,
+        agent_id: Optional[str] = None,
+        team_id: Optional[str] = None,
+        user_id: Optional[str] = None,
+    ) -> str:
+        """Creates memories from multiple messages and adds them to the memory db."""
+        self.set_log_level()
+
+        if self.db is None:
+            log_warning("MemoryDb not provided.")
+            return "Please provide a db to store memories"
+
+        if not messages and not message:
+            raise ValueError("You must provide either a message or a list of messages")
+
+        if message:
+            messages = [Message(role="user", content=message)]
+
+        if not messages or not isinstance(messages, list):
+            raise ValueError("Invalid messages list")
+
+        if user_id is None:
+            user_id = "default"
+
+        memories = self.read_from_db(user_id=user_id)
+        if memories is None:
+            memories = {}
+
+        existing_memories = memories.get(user_id, [])  # type: ignore
+        existing_memories = [{"memory_id": memory.memory_id, "memory": memory.memory} for memory in existing_memories]
+
+        response = await self.acreate_or_update_memories(  # type: ignore
+            messages=messages,
+            existing_memories=existing_memories,
+            user_id=user_id,
+            agent_id=agent_id,
+            team_id=team_id,
+            db=self.db,
+            update_memories=self.update_memories,
+            add_memories=self.add_memories,
+        )
+
+        # We refresh from the DB
+        self.read_from_db(user_id=user_id)
+
+        return response
+
+    def update_memory_task(self, task: str, user_id: Optional[str] = None) -> str:
+        """Updates the memory with a task"""
+
+        if not self.db:
+            log_warning("MemoryDb not provided.")
+            return "Please provide a db to store memories"
+
+        if user_id is None:
+            user_id = "default"
+
+        memories = self.read_from_db(user_id=user_id)
+        if memories is None:
+            memories = {}
+
+        existing_memories = memories.get(user_id, [])  # type: ignore
+        existing_memories = [{"memory_id": memory.memory_id, "memory": memory.memory} for memory in existing_memories]
+        # The memory manager updates the DB directly
+        response = self.run_memory_task(  # type: ignore
+            task=task,
+            existing_memories=existing_memories,
+            user_id=user_id,
+            db=self.db,
+            delete_memories=self.delete_memories,
+            update_memories=self.update_memories,
+            add_memories=self.add_memories,
+            clear_memories=self.clear_memories,
+        )
+
+        # We refresh from the DB
+        self.read_from_db(user_id=user_id)
+
+        return response
+
+    async def aupdate_memory_task(self, task: str, user_id: Optional[str] = None) -> str:
+        """Updates the memory with a task"""
+        self.set_log_level()
+
+        if not self.db:
+            log_warning("MemoryDb not provided.")
+            return "Please provide a db to store memories"
+
+        if user_id is None:
+            user_id = "default"
+
+        memories = self.read_from_db(user_id=user_id)
+        if memories is None:
+            memories = {}
+
+        existing_memories = memories.get(user_id, [])  # type: ignore
+        existing_memories = [{"memory_id": memory.memory_id, "memory": memory.memory} for memory in existing_memories]
+        # The memory manager updates the DB directly
+        response = await self.arun_memory_task(  # type: ignore
+            task=task,
+            existing_memories=existing_memories,
+            user_id=user_id,
+            db=self.db,
+            delete_memories=self.delete_memories,
+            update_memories=self.update_memories,
+            add_memories=self.add_memories,
+            clear_memories=self.clear_memories,
+        )
+
+        # We refresh from the DB
+        self.read_from_db(user_id=user_id)
+
+        return response
+
+    # -*- Memory Db Functions
+    def _upsert_db_memory(self, memory: UserMemory) -> str:
+        """Use this function to add a memory to the database."""
         try:
-            if self.db:
-                self.db.upsert_memory(
-                    MemoryRow(user_id=self.user_id, memory=Memory(memory=memory, input=self.input_message).to_dict())
-                )
+            if not self.db:
+                raise ValueError("Memory db not initialized")
+            self.db.upsert_user_memory(memory=memory)
             return "Memory added successfully"
         except Exception as e:
-            logger.warning(f"Error storing memory in db: {e}")
+            log_warning(f"Error storing memory in db: {e}")
             return f"Error adding memory: {e}"
 
-    def delete_memory(self, id: str) -> str:
-        """Use this function to delete a memory from the database.
-        Args:
-            id (str): The id of the memory to be deleted.
-        Returns:
-            str: A message indicating if the memory was deleted successfully or not.
-        """
+    def _delete_db_memory(self, memory_id: str) -> str:
+        """Use this function to delete a memory from the database."""
         try:
-            if self.db:
-                self.db.delete_memory(id=id)
+            if not self.db:
+                raise ValueError("Memory db not initialized")
+            self.db.delete_user_memory(memory_id=memory_id)
             return "Memory deleted successfully"
         except Exception as e:
-            logger.warning(f"Error deleting memory in db: {e}")
+            log_warning(f"Error deleting memory in db: {e}")
             return f"Error deleting memory: {e}"
 
-    def update_memory(self, id: str, memory: str) -> str:
-        """Use this function to update a memory in the database.
+    # -*- Utility Functions
+    def search_user_memories(
+        self,
+        query: Optional[str] = None,
+        limit: Optional[int] = None,
+        retrieval_method: Optional[Literal["last_n", "first_n", "agentic"]] = None,
+        user_id: Optional[str] = None,
+    ) -> List[UserMemory]:
+        """Search through user memories using the specified retrieval method.
+
         Args:
-            id (str): The id of the memory to be updated.
-            memory (str): The updated memory.
+            query: The search query for agentic search. Required if retrieval_method is "agentic".
+            limit: Maximum number of memories to return. Defaults to self.retrieval_limit if not specified. Optional.
+            retrieval_method: The method to use for retrieving memories. Defaults to self.retrieval if not specified.
+                - "last_n": Return the most recent memories
+                - "first_n": Return the oldest memories
+                - "agentic": Return memories most similar to the query, but using an agentic approach
+            user_id: The user to search for. Optional.
+
         Returns:
-            str: A message indicating if the memory was updated successfully or not.
+            A list of UserMemory objects matching the search criteria.
         """
-        try:
-            if self.db:
-                self.db.upsert_memory(
-                    MemoryRow(
-                        id=id, user_id=self.user_id, memory=Memory(memory=memory, input=self.input_message).to_dict()
-                    )
-                )
-            return "Memory updated successfully"
-        except Exception as e:
-            logger.warning(f"Error updating memory in db: {e}")
-            return f"Error updating memory: {e}"
 
-    def clear_memory(self) -> str:
-        """Use this function to clear all memories from the database.
+        if user_id is None:
+            user_id = "default"
+
+        self.set_log_level()
+
+        memories = self.read_from_db(user_id=user_id)
+        if memories is None:
+            memories = {}
+
+        if not memories:
+            return []
+
+        # Use default retrieval method if not specified
+        retrieval_method = retrieval_method
+        # Use default limit if not specified
+        limit = limit
+
+        # Handle different retrieval methods
+        if retrieval_method == "agentic":
+            if not query:
+                raise ValueError("Query is required for agentic search")
+
+            return self._search_user_memories_agentic(user_id=user_id, query=query, limit=limit)
+
+        elif retrieval_method == "first_n":
+            return self._get_first_n_memories(user_id=user_id, limit=limit)
+
+        else:  # Default to last_n
+            return self._get_last_n_memories(user_id=user_id, limit=limit)
+
+    def _get_response_format(self) -> Union[Dict[str, Any], Type[BaseModel]]:
+        model = self.get_model()
+        if model.supports_native_structured_outputs:
+            return MemorySearchResponse
+
+        elif model.supports_json_schema_outputs:
+            return {
+                "type": "json_schema",
+                "json_schema": {
+                    "name": MemorySearchResponse.__name__,
+                    "schema": MemorySearchResponse.model_json_schema(),
+                },
+            }
+        else:
+            return {"type": "json_object"}
+
+    def _search_user_memories_agentic(self, user_id: str, query: str, limit: Optional[int] = None) -> List[UserMemory]:
+        """Search through user memories using agentic search."""
+        memories = self.read_from_db(user_id=user_id)
+        if memories is None:
+            memories = {}
+
+        if not memories:
+            return []
+
+        model = self.get_model()
+
+        response_format = self._get_response_format()
+
+        log_debug("Searching for memories", center=True)
+
+        # Get all memories as a list
+        user_memories: List[UserMemory] = memories[user_id]
+        system_message_str = "Your task is to search through user memories and return the IDs of the memories that are related to the query.\n"
+        system_message_str += "\n<user_memories>\n"
+        for memory in user_memories:
+            system_message_str += f"ID: {memory.memory_id}\n"
+            system_message_str += f"Memory: {memory.memory}\n"
+            if memory.topics:
+                system_message_str += f"Topics: {','.join(memory.topics)}\n"
+            system_message_str += "\n"
+        system_message_str = system_message_str.strip()
+        system_message_str += "\n</user_memories>\n\n"
+        system_message_str += "REMEMBER: Only return the IDs of the memories that are related to the query."
+
+        if response_format == {"type": "json_object"}:
+            system_message_str += "\n" + get_json_output_prompt(MemorySearchResponse)  # type: ignore
+
+        messages_for_model = [
+            Message(role="system", content=system_message_str),
+            Message(
+                role="user",
+                content=f"Return the IDs of the memories related to the following query: {query}",
+            ),
+        ]
+
+        # Generate a response from the Model (includes running function calls)
+        response = model.response(messages=messages_for_model, response_format=response_format)
+        log_debug("Search for memories complete", center=True)
+
+        memory_search: Optional[MemorySearchResponse] = None
+        # If the model natively supports structured outputs, the parsed value is already in the structured format
+        if (
+            model.supports_native_structured_outputs
+            and response.parsed is not None
+            and isinstance(response.parsed, MemorySearchResponse)
+        ):
+            memory_search = response.parsed
+
+        # Otherwise convert the response to the structured format
+        if isinstance(response.content, str):
+            try:
+                memory_search = parse_response_model_str(response.content, MemorySearchResponse)  # type: ignore
+
+                # Update RunOutput
+                if memory_search is None:
+                    log_warning("Failed to convert memory_search response to MemorySearchResponse")
+                    return []
+            except Exception as e:
+                log_warning(f"Failed to convert memory_search response to MemorySearchResponse: {e}")
+                return []
+
+        memories_to_return = []
+        if memory_search:
+            for memory_id in memory_search.memory_ids:
+                for memory in user_memories:
+                    if memory.memory_id == memory_id:
+                        memories_to_return.append(memory)
+        return memories_to_return[:limit]
+
+    def _get_last_n_memories(self, user_id: str, limit: Optional[int] = None) -> List[UserMemory]:
+        """Get the most recent user memories.
+
+        Args:
+            limit: Maximum number of memories to return.
 
         Returns:
-            str: A message indicating if the memory was cleared successfully or not.
+            A list of the most recent UserMemory objects.
         """
-        try:
-            if self.db:
-                self.db.clear()
-            return "Memory cleared successfully"
-        except Exception as e:
-            logger.warning(f"Error clearing memory in db: {e}")
-            return f"Error clearing memory: {e}"
+        memories = self.read_from_db(user_id=user_id)
+        if memories is None:
+            memories = {}
+
+        memories_list = memories.get(user_id, [])
+
+        # Sort memories by updated_at timestamp if available
+        if memories_list:
+            # Sort memories by updated_at timestamp (newest first)
+            # If updated_at is None, place at the beginning of the list
+            sorted_memories_list = sorted(
+                memories_list,
+                key=lambda memory: memory.updated_at or datetime.min,
+            )
+        else:
+            sorted_memories_list = []
+
+        if limit is not None and limit > 0:
+            sorted_memories_list = sorted_memories_list[-limit:]
+
+        return sorted_memories_list
+
+    def _get_first_n_memories(self, user_id: str, limit: Optional[int] = None) -> List[UserMemory]:
+        """Get the oldest user memories.
+
+        Args:
+            limit: Maximum number of memories to return.
+
+        Returns:
+            A list of the oldest UserMemory objects.
+        """
+        memories = self.read_from_db(user_id=user_id)
+        if memories is None:
+            memories = {}
+
+        memories_list = memories.get(user_id, [])
+        # Sort memories by updated_at timestamp if available
+        if memories_list:
+            # Sort memories by updated_at timestamp (oldest first)
+            # If updated_at is None, place at the end of the list
+            sorted_memories_list = sorted(
+                memories_list,
+                key=lambda memory: memory.updated_at or datetime.max,
+            )
+
+        else:
+            sorted_memories_list = []
+
+        if limit is not None and limit > 0:
+            sorted_memories_list = sorted_memories_list[:limit]
+
+        return sorted_memories_list
+
+    # --Memory Manager Functions--
+    def determine_tools_for_model(self, tools: List[Callable]) -> None:
+        # Have to reset each time, because of different user IDs
+        self._tools_for_model = []
+        self._functions_for_model = {}
+
+        for tool in tools:
+            try:
+                function_name = tool.__name__
+                if function_name not in self._functions_for_model:
+                    func = Function.from_callable(tool, strict=True)  # type: ignore
+                    func.strict = True
+                    self._functions_for_model[func.name] = func
+                    self._tools_for_model.append({"type": "function", "function": func.to_dict()})
+                    log_debug(f"Added function {func.name}")
+            except Exception as e:
+                log_warning(f"Could not add function {tool}: {e}")
+
+    def get_system_message(
+        self,
+        existing_memories: Optional[List[Dict[str, Any]]] = None,
+        enable_delete_memory: bool = True,
+        enable_clear_memory: bool = True,
+        enable_update_memory: bool = True,
+        enable_add_memory: bool = True,
+    ) -> Message:
+        if self.system_message is not None:
+            return Message(role="system", content=self.system_message)
+
+        memory_capture_instructions = self.memory_capture_instructions or dedent("""\
+            Memories should include details that could personalize ongoing interactions with the user, such as:
+              - Personal facts: name, age, occupation, location, interests, preferences, etc.
+              - Significant life events or experiences shared by the user
+              - Important context about the user's current situation, challenges or goals
+              - What the user likes or dislikes, their opinions, beliefs, values, etc.
+              - Any other details that provide valuable insights into the user's personality, perspective or needs\
+        """)
 
-    def get_system_message(self) -> Message:
         # -*- Return a system message for the memory manager
         system_prompt_lines = [
-            "Your task is to generate a concise memory for the user's message. "
-            "Create a memory that captures the key information provided by the user, as if you were storing it for future reference. "
-            "The memory should be a brief, third-person statement that encapsulates the most important aspect of the user's input, without adding any extraneous details. "
-            "This memory will be used to enhance the user's experience in subsequent conversations.",
-            "You will also be provided with a list of existing memories. You may:",
-            "  1. Add a new memory using the `add_memory` tool.",
-            "  2. Update a memory using the `update_memory` tool.",
-            "  3. Delete a memory using the `delete_memory` tool.",
-            "  4. Clear all memories using the `clear_memory` tool. Use this with extreme caution, as it will remove all memories from the database.",
+            "You are a MemoryConnector that is responsible for manging key information about the user. "
+            "You will be provided with a criteria for memories to capture in the <memories_to_capture> section and a list of existing memories in the <existing_memories> section.",
+            "",
+            "## When to add or update memories",
+            "- Your first task is to decide if a memory needs to be added, updated, or deleted based on the user's message OR if no changes are needed.",
+            "- If the user's message meets the criteria in the <memories_to_capture> section and that information is not already captured in the <existing_memories> section, you should capture it as a memory.",
+            "- If the users messages does not meet the criteria in the <memories_to_capture> section, no memory updates are needed.",
+            "- If the existing memories in the <existing_memories> section capture all relevant information, no memory updates are needed.",
+            "",
+            "## How to add or update memories",
+            "- If you decide to add a new memory, create memories that captures key information, as if you were storing it for future reference.",
+            "- Memories should be a brief, third-person statements that encapsulate the most important aspect of the user's input, without adding any extraneous information.",
+            "  - Example: If the user's message is 'I'm going to the gym', a memory could be `John Doe goes to the gym regularly`.",
+            "  - Example: If the user's message is 'My name is John Doe', a memory could be `User's name is John Doe`.",
+            "- Don't make a single memory too long or complex, create multiple memories if needed to capture all the information.",
+            "- Don't repeat the same information in multiple memories. Rather update existing memories if needed.",
+            "- If a user asks for a memory to be updated or forgotten, remove all reference to the information that should be forgotten. Don't say 'The user used to like ...`",
+            "- When updating a memory, append the existing memory with new information rather than completely overwriting it.",
+            "- When a user's preferences change, update the relevant memories to reflect the new preferences but also capture what the user's preferences used to be and what has changed.",
+            "",
+            "## Criteria for creating memories",
+            "Use the following criteria to determine if a user's message should be captured as a memory.",
+            "",
+            "<memories_to_capture>",
+            memory_capture_instructions,
+            "</memories_to_capture>",
+            "",
+            "## Updating memories",
+            "You will also be provided with a list of existing memories in the <existing_memories> section. You can:",
+            "  1. Decide to make no changes.",
+        ]
+        if enable_add_memory:
+            system_prompt_lines.append("  2. Decide to add a new memory, using the `add_memory` tool.")
+        if enable_update_memory:
+            system_prompt_lines.append("  3. Decide to update an existing memory, using the `update_memory` tool.")
+        if enable_delete_memory:
+            system_prompt_lines.append("  4. Decide to delete an existing memory, using the `delete_memory` tool.")
+        if enable_clear_memory:
+            system_prompt_lines.append("  5. Decide to clear all memories, using the `clear_memory` tool.")
+
+        system_prompt_lines += [
+            "You can call multiple tools in a single response if needed. ",
+            "Only add or update memories if it is necessary to capture key information provided by the user.",
         ]
-        existing_memories = self.get_existing_memories()
+
         if existing_memories and len(existing_memories) > 0:
-            system_prompt_lines.extend(
-                [
-                    "\nExisting memories:",
-                    "<existing_memories>\n"
-                    + "\n".join([f"  - id: {m.id} | memory: {m.memory}" for m in existing_memories])
-                    + "\n</existing_memories>",
-                ]
-            )
+            system_prompt_lines.append("\n<existing_memories>")
+            for existing_memory in existing_memories:
+                system_prompt_lines.append(f"ID: {existing_memory['memory_id']}")
+                system_prompt_lines.append(f"Memory: {existing_memory['memory']}")
+                system_prompt_lines.append("")
+            system_prompt_lines.append("</existing_memories>")
+
+        if self.additional_instructions:
+            system_prompt_lines.append(self.additional_instructions)
+
         return Message(role="system", content="\n".join(system_prompt_lines))
 
-    def run(
+    def create_or_update_memories(
         self,
-        message: Optional[str] = None,
-        **kwargs: Any,
-    ) -> Optional[str]:
-        log_debug("*********** MemoryManager Start ***********")
+        messages: List[Message],
+        existing_memories: List[Dict[str, Any]],
+        user_id: str,
+        db: BaseDb,
+        agent_id: Optional[str] = None,
+        team_id: Optional[str] = None,
+        update_memories: bool = True,
+        add_memories: bool = True,
+    ) -> str:
+        if self.model is None:
+            log_error("No model provided for memory manager")
+            return "No model provided for memory manager"
+
+        log_debug("MemoryManager Start", center=True)
+
+        if len(messages) == 1:
+            input_string = messages[0].get_content_string()
+        else:
+            input_string = f"{', '.join([m.get_content_string() for m in messages if m.role == 'user' and m.content])}"
 
+        model_copy = deepcopy(self.model)
         # Update the Model (set defaults, add logit etc.)
-        self.update_model()
-        self.determine_tools_for_model()
+        self.determine_tools_for_model(
+            self._get_db_tools(
+                user_id,
+                db,
+                input_string,
+                agent_id=agent_id,
+                team_id=team_id,
+                enable_add_memory=add_memories,
+                enable_update_memory=update_memories,
+                enable_delete_memory=False,
+                enable_clear_memory=False,
+            ),
+        )
 
         # Prepare the List of messages to send to the Model
-        messages_for_model: List[Message] = [self.get_system_message()]
+        messages_for_model: List[Message] = [
+            self.get_system_message(
+                existing_memories=existing_memories,
+                enable_update_memory=update_memories,
+                enable_add_memory=add_memories,
+                enable_delete_memory=False,
+                enable_clear_memory=False,
+            ),
+            *messages,
+        ]
+
+        # Generate a response from the Model (includes running function calls)
+        response = model_copy.response(
+            messages=messages_for_model, tools=self._tools_for_model, functions=self._functions_for_model
+        )
+
+        if response.tool_calls is not None and len(response.tool_calls) > 0:
+            self.memories_updated = True
+        log_debug("MemoryManager End", center=True)
+
+        return response.content or "No response from model"
+
+    async def acreate_or_update_memories(
+        self,
+        messages: List[Message],
+        existing_memories: List[Dict[str, Any]],
+        user_id: str,
+        db: BaseDb,
+        agent_id: Optional[str] = None,
+        team_id: Optional[str] = None,
+        update_memories: bool = True,
+        add_memories: bool = True,
+    ) -> str:
+        if self.model is None:
+            log_error("No model provided for memory manager")
+            return "No model provided for memory manager"
 
-        # Add the user prompt message
-        user_prompt_message = Message(role="user", content=message, **kwargs) if message else None
-        if user_prompt_message is not None:
-            messages_for_model += [user_prompt_message]
+        log_debug("MemoryManager Start", center=True)
 
-        # Set input message added with the memory
-        self.input_message = message
+        if len(messages) == 1:
+            input_string = messages[0].get_content_string()
+        else:
+            input_string = f"{', '.join([m.get_content_string() for m in messages if m.role == 'user' and m.content])}"
+
+        model_copy = deepcopy(self.model)
+        # Update the Model (set defaults, add logit etc.)
+        self.determine_tools_for_model(
+            self._get_db_tools(
+                user_id,
+                db,
+                input_string,
+                agent_id=agent_id,
+                team_id=team_id,
+                enable_add_memory=add_memories,
+                enable_update_memory=update_memories,
+                enable_delete_memory=False,
+                enable_clear_memory=False,
+            ),
+        )
+
+        # Prepare the List of messages to send to the Model
+        messages_for_model: List[Message] = [
+            self.get_system_message(
+                existing_memories=existing_memories,
+                enable_update_memory=update_memories,
+                enable_add_memory=add_memories,
+                enable_delete_memory=False,
+                enable_clear_memory=False,
+            ),
+            *messages,
+        ]
 
         # Generate a response from the Model (includes running function calls)
-        self.model = cast(Model, self.model)
-        response = self.model.response(
+        response = await model_copy.aresponse(
             messages=messages_for_model, tools=self._tools_for_model, functions=self._functions_for_model
         )
-        log_debug("*********** MemoryManager End ***********")
-        return response.content
 
-    async def arun(
+        if response.tool_calls is not None and len(response.tool_calls) > 0:
+            self.memories_updated = True
+        log_debug("MemoryManager End", center=True)
+
+        return response.content or "No response from model"
+
+    def run_memory_task(
         self,
-        message: Optional[str] = None,
-        **kwargs: Any,
-    ) -> Optional[str]:
-        log_debug("*********** Async MemoryManager Start ***********")
+        task: str,
+        existing_memories: List[Dict[str, Any]],
+        user_id: str,
+        db: BaseDb,
+        delete_memories: bool = True,
+        update_memories: bool = True,
+        add_memories: bool = True,
+        clear_memories: bool = True,
+    ) -> str:
+        if self.model is None:
+            log_error("No model provided for memory manager")
+            return "No model provided for memory manager"
 
+        log_debug("MemoryManager Start", center=True)
+
+        model_copy = deepcopy(self.model)
         # Update the Model (set defaults, add logit etc.)
-        self.update_model()
+        self.determine_tools_for_model(
+            self._get_db_tools(
+                user_id,
+                db,
+                task,
+                enable_delete_memory=delete_memories,
+                enable_clear_memory=clear_memories,
+                enable_update_memory=update_memories,
+                enable_add_memory=add_memories,
+            ),
+        )
 
         # Prepare the List of messages to send to the Model
-        messages_for_model: List[Message] = [self.get_system_message()]
-        # Add the user prompt message
-        user_prompt_message = Message(role="user", content=message, **kwargs) if message else None
-        if user_prompt_message is not None:
-            messages_for_model += [user_prompt_message]
+        messages_for_model: List[Message] = [
+            self.get_system_message(
+                existing_memories,
+                enable_delete_memory=delete_memories,
+                enable_clear_memory=clear_memories,
+                enable_update_memory=update_memories,
+                enable_add_memory=add_memories,
+            ),
+            # For models that require a non-system message
+            Message(role="user", content=task),
+        ]
+
+        # Generate a response from the Model (includes running function calls)
+        response = model_copy.response(
+            messages=messages_for_model, tools=self._tools_for_model, functions=self._functions_for_model
+        )
+
+        if response.tool_calls is not None and len(response.tool_calls) > 0:
+            self.memories_updated = True
+        log_debug("MemoryManager End", center=True)
+
+        return response.content or "No response from model"
+
+    async def arun_memory_task(
+        self,
+        task: str,
+        existing_memories: List[Dict[str, Any]],
+        user_id: str,
+        db: BaseDb,
+        delete_memories: bool = True,
+        clear_memories: bool = True,
+        update_memories: bool = True,
+        add_memories: bool = True,
+    ) -> str:
+        if self.model is None:
+            log_error("No model provided for memory manager")
+            return "No model provided for memory manager"
+
+        log_debug("MemoryManager Start", center=True)
+
+        model_copy = deepcopy(self.model)
+        # Update the Model (set defaults, add logit etc.)
+        self.determine_tools_for_model(
+            self._get_db_tools(
+                user_id,
+                db,
+                task,
+                enable_delete_memory=delete_memories,
+                enable_clear_memory=clear_memories,
+                enable_update_memory=update_memories,
+                enable_add_memory=add_memories,
+            ),
+        )
 
-        # Set input message added with the memory
-        self.input_message = message
+        # Prepare the List of messages to send to the Model
+        messages_for_model: List[Message] = [
+            self.get_system_message(
+                existing_memories,
+                enable_delete_memory=delete_memories,
+                enable_clear_memory=clear_memories,
+                enable_update_memory=update_memories,
+                enable_add_memory=add_memories,
+            ),
+            # For models that require a non-system message
+            Message(role="user", content=task),
+        ]
 
         # Generate a response from the Model (includes running function calls)
-        self.model = cast(Model, self.model)
-        response = await self.model.aresponse(
+        response = await model_copy.aresponse(
             messages=messages_for_model, tools=self._tools_for_model, functions=self._functions_for_model
         )
-        log_debug("*********** Async MemoryManager End ***********")
-        return response.content
+
+        if response.tool_calls is not None and len(response.tool_calls) > 0:
+            self.memories_updated = True
+        log_debug("MemoryManager End", center=True)
+
+        return response.content or "No response from model"
+
+    # -*- DB Functions
+    def _get_db_tools(
+        self,
+        user_id: str,
+        db: BaseDb,
+        input_string: str,
+        enable_add_memory: bool = True,
+        enable_update_memory: bool = True,
+        enable_delete_memory: bool = True,
+        enable_clear_memory: bool = True,
+        agent_id: Optional[str] = None,
+        team_id: Optional[str] = None,
+    ) -> List[Callable]:
+        def add_memory(memory: str, topics: Optional[List[str]] = None) -> str:
+            """Use this function to add a memory to the database.
+            Args:
+                memory (str): The memory to be added.
+                topics (Optional[List[str]]): The topics of the memory (e.g. ["name", "hobbies", "location"]).
+            Returns:
+                str: A message indicating if the memory was added successfully or not.
+            """
+            from uuid import uuid4
+
+            from agno.db.base import UserMemory
+
+            try:
+                memory_id = str(uuid4())
+                db.upsert_user_memory(
+                    UserMemory(
+                        memory_id=memory_id,
+                        user_id=user_id,
+                        agent_id=agent_id,
+                        team_id=team_id,
+                        memory=memory,
+                        topics=topics,
+                        input=input_string,
+                    )
+                )
+                log_debug(f"Memory added: {memory_id}")
+                return "Memory added successfully"
+            except Exception as e:
+                log_warning(f"Error storing memory in db: {e}")
+                return f"Error adding memory: {e}"
+
+        def update_memory(memory_id: str, memory: str, topics: Optional[List[str]] = None) -> str:
+            """Use this function to update an existing memory in the database.
+            Args:
+                memory_id (str): The id of the memory to be updated.
+                memory (str): The updated memory.
+                topics (Optional[List[str]]): The topics of the memory (e.g. ["name", "hobbies", "location"]).
+            Returns:
+                str: A message indicating if the memory was updated successfully or not.
+            """
+            from agno.db.base import UserMemory
+
+            try:
+                db.upsert_user_memory(
+                    UserMemory(
+                        memory_id=memory_id,
+                        memory=memory,
+                        topics=topics,
+                        input=input_string,
+                    )
+                )
+                log_debug("Memory updated")
+                return "Memory updated successfully"
+            except Exception as e:
+                log_warning(f"Error storing memory in db: {e}")
+                return f"Error adding memory: {e}"
+
+        def delete_memory(memory_id: str) -> str:
+            """Use this function to delete a single memory from the database.
+            Args:
+                memory_id (str): The id of the memory to be deleted.
+            Returns:
+                str: A message indicating if the memory was deleted successfully or not.
+            """
+            try:
+                db.delete_user_memory(memory_id=memory_id)
+                log_debug("Memory deleted")
+                return "Memory deleted successfully"
+            except Exception as e:
+                log_warning(f"Error deleting memory in db: {e}")
+                return f"Error deleting memory: {e}"
+
+        def clear_memory() -> str:
+            """Use this function to remove all (or clear all) memories from the database.
+
+            Returns:
+                str: A message indicating if the memory was cleared successfully or not.
+            """
+            db.clear_memories()
+            log_debug("Memory cleared")
+            return "Memory cleared successfully"
+
+        functions: List[Callable] = []
+        if enable_add_memory:
+            functions.append(add_memory)
+        if enable_update_memory:
+            functions.append(update_memory)
+        if enable_delete_memory:
+            functions.append(delete_memory)
+        if enable_clear_memory:
+            functions.append(clear_memory)
+        return functions