letta-nightly 0.4.1.dev20241007104134__py3-none-any.whl → 0.4.1.dev20241009104130__py3-none-any.whl

letta/providers.py

@@ -1,8 +1,13 @@
 from typing import List, Optional
 
-from pydantic import BaseModel, Field
+from pydantic import BaseModel, Field, model_validator
 
 from letta.constants import LLM_MAX_TOKENS
+from letta.llm_api.azure_openai import (
+    get_azure_chat_completions_endpoint,
+    get_azure_embeddings_endpoint,
+)
+from letta.llm_api.azure_openai_constants import AZURE_MODEL_TO_CONTEXT_LENGTH
 from letta.schemas.embedding_config import EmbeddingConfig
 from letta.schemas.llm_config import LLMConfig
 
@@ -122,34 +127,64 @@ class OllamaProvider(OpenAIProvider):
         response = requests.post(f"{self.base_url}/api/show", json={"name": model_name, "verbose": True})
         response_json = response.json()
 
-        # thank you vLLM: https://github.com/vllm-project/vllm/blob/main/vllm/config.py#L1675
-        possible_keys = [
-            # OPT
-            "max_position_embeddings",
-            # GPT-2
-            "n_positions",
-            # MPT
-            "max_seq_len",
-            # ChatGLM2
-            "seq_length",
-            # Command-R
-            "model_max_length",
-            # Others
-            "max_sequence_length",
-            "max_seq_length",
-            "seq_len",
-        ]
-
+        ## thank you vLLM: https://github.com/vllm-project/vllm/blob/main/vllm/config.py#L1675
+        # possible_keys = [
+        #    # OPT
+        #    "max_position_embeddings",
+        #    # GPT-2
+        #    "n_positions",
+        #    # MPT
+        #    "max_seq_len",
+        #    # ChatGLM2
+        #    "seq_length",
+        #    # Command-R
+        #    "model_max_length",
+        #    # Others
+        #    "max_sequence_length",
+        #    "max_seq_length",
+        #    "seq_len",
+        # ]
         # max_position_embeddings
         # parse model cards: nous, dolphon, llama
         for key, value in response_json["model_info"].items():
-            if "context_window" in key:
+            if "context_length" in key:
+                return value
+        return None
+
+    def get_model_embedding_dim(self, model_name: str):
+        import requests
+
+        response = requests.post(f"{self.base_url}/api/show", json={"name": model_name, "verbose": True})
+        response_json = response.json()
+        for key, value in response_json["model_info"].items():
+            if "embedding_length" in key:
                 return value
         return None
 
     def list_embedding_models(self) -> List[EmbeddingConfig]:
-        # TODO: filter embedding models
-        return []
+        # https://github.com/ollama/ollama/blob/main/docs/api.md#list-local-models
+        import requests
+
+        response = requests.get(f"{self.base_url}/api/tags")
+        if response.status_code != 200:
+            raise Exception(f"Failed to list Ollama models: {response.text}")
+        response_json = response.json()
+
+        configs = []
+        for model in response_json["models"]:
+            embedding_dim = self.get_model_embedding_dim(model["name"])
+            if not embedding_dim:
+                continue
+            configs.append(
+                EmbeddingConfig(
+                    embedding_model=model["name"],
+                    embedding_endpoint_type="ollama",
+                    embedding_endpoint=self.base_url,
+                    embedding_dim=embedding_dim,
+                    embedding_chunk_size=300,
+                )
+            )
+        return configs
 
 
 class GroqProvider(OpenAIProvider):
@@ -182,20 +217,21 @@ class GroqProvider(OpenAIProvider):
 class GoogleAIProvider(Provider):
     # gemini
     api_key: str = Field(..., description="API key for the Google AI API.")
-    service_endpoint: str = "generativelanguage"
     base_url: str = "https://generativelanguage.googleapis.com"
 
     def list_llm_models(self):
         from letta.llm_api.google_ai import google_ai_get_model_list
 
-        # TODO: use base_url instead
-        model_options = google_ai_get_model_list(service_endpoint=self.service_endpoint, api_key=self.api_key)
+        model_options = google_ai_get_model_list(base_url=self.base_url, api_key=self.api_key)
+        # filter by 'generateContent' models
+        model_options = [mo for mo in model_options if "generateContent" in mo["supportedGenerationMethods"]]
         model_options = [str(m["name"]) for m in model_options]
+
+        # filter by model names
         model_options = [mo[len("models/") :] if mo.startswith("models/") else mo for mo in model_options]
+
         # TODO remove manual filtering for gemini-pro
         model_options = [mo for mo in model_options if str(mo).startswith("gemini") and "-pro" in str(mo)]
-        # TODO: add context windows
-        # model_options = ["gemini-pro"]
 
         configs = []
         for model in model_options:
@@ -210,17 +246,94 @@ class GoogleAIProvider(Provider):
         return configs
 
     def list_embedding_models(self):
-        return []
+        from letta.llm_api.google_ai import google_ai_get_model_list
+
+        # TODO: use base_url instead
+        model_options = google_ai_get_model_list(base_url=self.base_url, api_key=self.api_key)
+        # filter by 'generateContent' models
+        model_options = [mo for mo in model_options if "embedContent" in mo["supportedGenerationMethods"]]
+        model_options = [str(m["name"]) for m in model_options]
+        model_options = [mo[len("models/") :] if mo.startswith("models/") else mo for mo in model_options]
+
+        configs = []
+        for model in model_options:
+            configs.append(
+                EmbeddingConfig(
+                    embedding_model=model,
+                    embedding_endpoint_type="google_ai",
+                    embedding_endpoint=self.base_url,
+                    embedding_dim=768,
+                    embedding_chunk_size=300,  # NOTE: max is 2048
+                )
+            )
+        return configs
 
     def get_model_context_window(self, model_name: str):
         from letta.llm_api.google_ai import google_ai_get_model_context_window
 
-        # TODO: use base_url instead
-        return google_ai_get_model_context_window(self.service_endpoint, self.api_key, model_name)
+        return google_ai_get_model_context_window(self.base_url, self.api_key, model_name)
 
 
 class AzureProvider(Provider):
-    pass
+    name: str = "azure"
+    latest_api_version: str = "2024-09-01-preview"  # https://learn.microsoft.com/en-us/azure/ai-services/openai/api-version-deprecation
+    base_url: str = Field(
+        ..., description="Base URL for the Azure API endpoint. This should be specific to your org, e.g. `https://letta.openai.azure.com`."
+    )
+    api_key: str = Field(..., description="API key for the Azure API.")
+    api_version: str = Field(latest_api_version, description="API version for the Azure API")
+
+    @model_validator(mode="before")
+    def set_default_api_version(cls, values):
+        """
+        This ensures that api_version is always set to the default if None is passed in.
+        """
+        if values.get("api_version") is None:
+            values["api_version"] = cls.model_fields["latest_api_version"].default
+        return values
+
+    def list_llm_models(self) -> List[LLMConfig]:
+        from letta.llm_api.azure_openai import (
+            azure_openai_get_chat_completion_model_list,
+        )
+
+        model_options = azure_openai_get_chat_completion_model_list(self.base_url, api_key=self.api_key, api_version=self.api_version)
+        configs = []
+        for model_option in model_options:
+            model_name = model_option["id"]
+            context_window_size = self.get_model_context_window(model_name)
+            model_endpoint = get_azure_chat_completions_endpoint(self.base_url, model_name, self.api_version)
+            configs.append(
+                LLMConfig(model=model_name, model_endpoint_type="azure", model_endpoint=model_endpoint, context_window=context_window_size)
+            )
+        return configs
+
+    def list_embedding_models(self) -> List[EmbeddingConfig]:
+        from letta.llm_api.azure_openai import azure_openai_get_embeddings_model_list
+
+        model_options = azure_openai_get_embeddings_model_list(
+            self.base_url, api_key=self.api_key, api_version=self.api_version, require_embedding_in_name=True
+        )
+        configs = []
+        for model_option in model_options:
+            model_name = model_option["id"]
+            model_endpoint = get_azure_embeddings_endpoint(self.base_url, model_name, self.api_version)
+            configs.append(
+                EmbeddingConfig(
+                    embedding_model=model_name,
+                    embedding_endpoint_type="azure",
+                    embedding_endpoint=model_endpoint,
+                    embedding_dim=768,
+                    embedding_chunk_size=300,  # NOTE: max is 2048
+                )
+            )
+        return configs
+
+    def get_model_context_window(self, model_name: str):
+        """
+        This is hardcoded for now, since there is no API endpoints to retrieve metadata for a model.
+        """
+        return AZURE_MODEL_TO_CONTEXT_LENGTH.get(model_name, 4096)
 
 
 class VLLMProvider(OpenAIProvider):

letta/schemas/agent.py

@@ -1,5 +1,6 @@
 import uuid
 from datetime import datetime
+from enum import Enum
 from typing import Dict, List, Optional, Union
 
 from pydantic import BaseModel, Field, field_validator
@@ -21,6 +22,15 @@ class BaseAgent(LettaBase, validate_assignment=True):
     user_id: Optional[str] = Field(None, description="The user id of the agent.")
 
 
+class AgentType(str, Enum):
+    """
+    Enum to represent the type of agent.
+    """
+
+    memgpt_agent = "memgpt_agent"
+    split_thread_agent = "split_thread_agent"
+
+
 class AgentState(BaseAgent):
     """
     Representation of an agent's state. This is the state of the agent at a given time, and is persisted in the DB backend. The state has all the information needed to recreate a persisted agent.
@@ -52,6 +62,9 @@ class AgentState(BaseAgent):
     # system prompt
     system: str = Field(..., description="The system prompt used by the agent.")
 
+    # agent configuration
+    agent_type: AgentType = Field(..., description="The type of agent.")
+
     # llm information
     llm_config: LLMConfig = Field(..., description="The LLM configuration used by the agent.")
     embedding_config: EmbeddingConfig = Field(..., description="The embedding configuration used by the agent.")
@@ -64,6 +77,7 @@ class CreateAgent(BaseAgent):
     memory: Optional[Memory] = Field(None, description="The in-context memory of the agent.")
     tools: Optional[List[str]] = Field(None, description="The tools used by the agent.")
     system: Optional[str] = Field(None, description="The system prompt used by the agent.")
+    agent_type: Optional[AgentType] = Field(None, description="The type of agent.")
     llm_config: Optional[LLMConfig] = Field(None, description="The LLM configuration used by the agent.")
     embedding_config: Optional[EmbeddingConfig] = Field(None, description="The embedding configuration used by the agent.")
 

letta/schemas/llm_config.py

@@ -11,7 +11,7 @@ class LLMConfig(BaseModel):
         model (str): The name of the LLM model.
         model_endpoint_type (str): The endpoint type for the model.
         model_endpoint (str): The endpoint for the model.
-        model_wrapper (str): The wrapper for the model.
+        model_wrapper (str): The wrapper for the model. This is used to wrap additional text around the input/output of the model. This is useful for text-to-text completions, such as the Completions API in OpenAI.
         context_window (int): The context window size for the model.
     """
 
@@ -34,7 +34,7 @@ class LLMConfig(BaseModel):
         "vllm",
         "hugging-face",
     ] = Field(..., description="The endpoint type for the model.")
-    model_endpoint: str = Field(..., description="The endpoint for the model.")
+    model_endpoint: Optional[str] = Field(None, description="The endpoint for the model.")
     model_wrapper: Optional[str] = Field(None, description="The wrapper for the model.")
     context_window: int = Field(..., description="The context window size for the model.")
 

letta/schemas/openai/chat_completion_response.py

@@ -74,6 +74,9 @@ class ChatCompletionResponse(BaseModel):
     object: Literal["chat.completion"] = "chat.completion"
     usage: UsageStatistics
 
+    def __str__(self):
+        return self.model_dump_json(indent=4)
+
 
 class FunctionCallDelta(BaseModel):
     # arguments: Optional[str] = None

letta/schemas/tool.py

@@ -93,7 +93,7 @@ class Tool(BaseTool):
         # append heartbeat (necessary for triggering another reasoning step after this tool call)
         json_schema["parameters"]["properties"]["request_heartbeat"] = {
             "type": "boolean",
-            "description": "Request an immediate heartbeat after function execution. Set to 'true' if you want to send a follow-up message or run a follow-up function.",
+            "description": "Request an immediate heartbeat after function execution. Set to `True` if you want to send a follow-up message or run a follow-up function.",
         }
         json_schema["parameters"]["required"].append("request_heartbeat")
 
@@ -128,7 +128,7 @@ class Tool(BaseTool):
         # append heartbeat (necessary for triggering another reasoning step after this tool call)
         json_schema["parameters"]["properties"]["request_heartbeat"] = {
             "type": "boolean",
-            "description": "Request an immediate heartbeat after function execution. Set to 'true' if you want to send a follow-up message or run a follow-up function.",
+            "description": "Request an immediate heartbeat after function execution. Set to `True` if you want to send a follow-up message or run a follow-up function.",
         }
         json_schema["parameters"]["required"].append("request_heartbeat")
 
@@ -161,7 +161,7 @@ class Tool(BaseTool):
         # append heartbeat (necessary for triggering another reasoning step after this tool call)
         json_schema["parameters"]["properties"]["request_heartbeat"] = {
             "type": "boolean",
-            "description": "Request an immediate heartbeat after function execution. Set to 'true' if you want to send a follow-up message or run a follow-up function.",
+            "description": "Request an immediate heartbeat after function execution. Set to `True` if you want to send a follow-up message or run a follow-up function.",
         }
         json_schema["parameters"]["required"].append("request_heartbeat")
 

letta/server/rest_api/admin/tools.py

@@ -26,7 +26,6 @@ class CreateToolResponse(BaseModel):
 
 
 def setup_tools_index_router(server: SyncServer, interface: QueuingInterface):
-    # get_current_user_with_server = partial(partial(get_current_user, server), password)
 
     @router.delete("/tools/{tool_name}", tags=["tools"])
     async def delete_tool(

letta/server/rest_api/app.py

@@ -5,8 +5,7 @@ from pathlib import Path
 from typing import Optional
 
 import uvicorn
-from fastapi import FastAPI, Request
-from fastapi.responses import JSONResponse
+from fastapi import FastAPI
 from starlette.middleware.cors import CORSMiddleware
 
 from letta.server.constants import REST_DEFAULT_PORT
@@ -84,21 +83,6 @@ def create_application() -> "FastAPI":
         allow_headers=["*"],
     )
 
-    @app.middleware("http")
-    async def set_current_user_middleware(request: Request, call_next):
-        user_id = request.headers.get("user_id")
-        if user_id:
-            try:
-                server.set_current_user(user_id)
-            except ValueError as e:
-                # Return an HTTP 401 Unauthorized response
-                # raise HTTPException(status_code=401, detail=str(e))
-                return JSONResponse(status_code=401, content={"detail": str(e)})
-        else:
-            server.set_current_user(None)
-        response = await call_next(request)
-        return response
-
     for route in v1_routes:
         app.include_router(route, prefix=API_PREFIX)
         # this gives undocumented routes for "latest" and bare api calls.

letta/server/rest_api/routers/openai/assistants/threads.py

@@ -1,7 +1,7 @@
 import uuid
-from typing import TYPE_CHECKING, List
+from typing import TYPE_CHECKING, List, Optional
 
-from fastapi import APIRouter, Body, Depends, HTTPException, Path, Query
+from fastapi import APIRouter, Body, Depends, Header, HTTPException, Path, Query
 
 from letta.constants import DEFAULT_PRESET
 from letta.schemas.agent import CreateAgent
@@ -43,11 +43,12 @@ router = APIRouter(prefix="/v1/threads", tags=["threads"])
 def create_thread(
     request: CreateThreadRequest = Body(...),
     server: SyncServer = Depends(get_letta_server),
+    user_id: Optional[str] = Header(None, alias="user_id"),  # Extract user_id from header, default to None if not present
 ):
     # TODO: use requests.description and requests.metadata fields
     # TODO: handle requests.file_ids and requests.tools
     # TODO: eventually allow request to override embedding/llm model
-    actor = server.get_current_user()
+    actor = server.get_user_or_default(user_id=user_id)
 
     print("Create thread/agent", request)
     # create a letta agent
@@ -67,8 +68,9 @@ def create_thread(
 def retrieve_thread(
     thread_id: str = Path(..., description="The unique identifier of the thread."),
     server: SyncServer = Depends(get_letta_server),
+    user_id: Optional[str] = Header(None, alias="user_id"),  # Extract user_id from header, default to None if not present
 ):
-    actor = server.get_current_user()
+    actor = server.get_user_or_default(user_id=user_id)
     agent = server.get_agent(user_id=actor.id, agent_id=thread_id)
     assert agent is not None
     return OpenAIThread(
@@ -100,8 +102,9 @@ def create_message(
     thread_id: str = Path(..., description="The unique identifier of the thread."),
     request: CreateMessageRequest = Body(...),
     server: SyncServer = Depends(get_letta_server),
+    user_id: Optional[str] = Header(None, alias="user_id"),  # Extract user_id from header, default to None if not present
 ):
-    actor = server.get_current_user()
+    actor = server.get_user_or_default(user_id=user_id)
     agent_id = thread_id
     # create message object
     message = Message(
@@ -143,8 +146,9 @@ def list_messages(
     after: str = Query(None, description="A cursor for use in pagination. `after` is an object ID that defines your place in the list."),
     before: str = Query(None, description="A cursor for use in pagination. `after` is an object ID that defines your place in the list."),
     server: SyncServer = Depends(get_letta_server),
+    user_id: Optional[str] = Header(None, alias="user_id"),  # Extract user_id from header, default to None if not present
 ):
-    actor = server.get_current_user()
+    actor = server.get_user_or_default(user_id)
     after_uuid = after if before else None
     before_uuid = before if before else None
     agent_id = thread_id
@@ -239,7 +243,6 @@ def create_run(
     request: CreateRunRequest = Body(...),
     server: SyncServer = Depends(get_letta_server),
 ):
-    server.get_current_user()
 
     # TODO: add request.instructions as a message?
     agent_id = thread_id

letta/server/rest_api/routers/openai/chat_completions/chat_completions.py

@@ -1,7 +1,7 @@
 import json
-from typing import TYPE_CHECKING
+from typing import TYPE_CHECKING, Optional
 
-from fastapi import APIRouter, Body, Depends, HTTPException
+from fastapi import APIRouter, Body, Depends, Header, HTTPException
 
 from letta.schemas.enums import MessageRole
 from letta.schemas.letta_message import FunctionCall, LettaMessage
@@ -30,12 +30,14 @@ router = APIRouter(prefix="/v1/chat/completions", tags=["chat_completions"])
 async def create_chat_completion(
     completion_request: ChatCompletionRequest = Body(...),
     server: "SyncServer" = Depends(get_letta_server),
+    user_id: Optional[str] = Header(None, alias="user_id"),  # Extract user_id from header, default to None if not present
 ):
     """Send a message to a Letta agent via a /chat/completions completion_request
     The bearer token will be used to identify the user.
     The 'user' field in the completion_request should be set to the agent ID.
     """
-    actor = server.get_current_user()
+    actor = server.get_user_or_default(user_id=user_id)
+
     agent_id = completion_request.user
     if agent_id is None:
         raise HTTPException(status_code=400, detail="Must pass agent_id in the 'user' field")

letta/server/rest_api/routers/v1/agents.py

@@ -2,7 +2,7 @@ import asyncio
 from datetime import datetime
 from typing import Dict, List, Optional, Union
 
-from fastapi import APIRouter, Body, Depends, HTTPException, Query, status
+from fastapi import APIRouter, Body, Depends, Header, HTTPException, Query, status
 from fastapi.responses import JSONResponse, StreamingResponse
 from starlette.responses import StreamingResponse
 
@@ -40,12 +40,13 @@ router = APIRouter(prefix="/agents", tags=["agents"])
 @router.get("/", response_model=List[AgentState], operation_id="list_agents")
 def list_agents(
     server: "SyncServer" = Depends(get_letta_server),
+    user_id: Optional[str] = Header(None, alias="user_id"),  # Extract user_id from header, default to None if not present
 ):
     """
     List all agents associated with a given user.
     This endpoint retrieves a list of all agents and their configurations associated with the specified user ID.
     """
-    actor = server.get_current_user()
+    actor = server.get_user_or_default(user_id=user_id)
 
     return server.list_agents(user_id=actor.id)
 
@@ -54,11 +55,12 @@ def list_agents(
 def create_agent(
     agent: CreateAgent = Body(...),
     server: "SyncServer" = Depends(get_letta_server),
+    user_id: Optional[str] = Header(None, alias="user_id"),  # Extract user_id from header, default to None if not present
 ):
     """
     Create a new agent with the specified configuration.
     """
-    actor = server.get_current_user()
+    actor = server.get_user_or_default(user_id=user_id)
     agent.user_id = actor.id
     # TODO: sarah make general
     # TODO: eventually remove this
@@ -74,9 +76,10 @@ def update_agent(
     agent_id: str,
     update_agent: UpdateAgentState = Body(...),
     server: "SyncServer" = Depends(get_letta_server),
+    user_id: Optional[str] = Header(None, alias="user_id"),  # Extract user_id from header, default to None if not present
 ):
     """Update an exsiting agent"""
-    actor = server.get_current_user()
+    actor = server.get_user_or_default(user_id=user_id)
 
     update_agent.id = agent_id
     return server.update_agent(update_agent, user_id=actor.id)
@@ -86,11 +89,12 @@ def update_agent(
 def get_agent_state(
     agent_id: str,
     server: "SyncServer" = Depends(get_letta_server),
+    user_id: Optional[str] = Header(None, alias="user_id"),  # Extract user_id from header, default to None if not present
 ):
     """
     Get the state of the agent.
     """
-    actor = server.get_current_user()
+    actor = server.get_user_or_default(user_id=user_id)
 
     if not server.ms.get_agent(user_id=actor.id, agent_id=agent_id):
         # agent does not exist
@@ -103,11 +107,12 @@ def get_agent_state(
 def delete_agent(
     agent_id: str,
     server: "SyncServer" = Depends(get_letta_server),
+    user_id: Optional[str] = Header(None, alias="user_id"),  # Extract user_id from header, default to None if not present
 ):
     """
     Delete an agent.
     """
-    actor = server.get_current_user()
+    actor = server.get_user_or_default(user_id=user_id)
 
     return server.delete_agent(user_id=actor.id, agent_id=agent_id)
 
@@ -120,7 +125,6 @@ def get_agent_sources(
     """
     Get the sources associated with an agent.
     """
-    server.get_current_user()
 
     return server.list_attached_sources(agent_id)
 
@@ -155,12 +159,13 @@ def update_agent_memory(
     agent_id: str,
     request: Dict = Body(...),
     server: "SyncServer" = Depends(get_letta_server),
+    user_id: Optional[str] = Header(None, alias="user_id"),  # Extract user_id from header, default to None if not present
 ):
     """
     Update the core memory of a specific agent.
     This endpoint accepts new memory contents (human and persona) and updates the core memory of the agent identified by the user ID and agent ID.
     """
-    actor = server.get_current_user()
+    actor = server.get_user_or_default(user_id=user_id)
 
     memory = server.update_agent_core_memory(user_id=actor.id, agent_id=agent_id, new_memory_contents=request)
     return memory
@@ -197,11 +202,12 @@ def get_agent_archival_memory(
     after: Optional[int] = Query(None, description="Unique ID of the memory to start the query range at."),
     before: Optional[int] = Query(None, description="Unique ID of the memory to end the query range at."),
     limit: Optional[int] = Query(None, description="How many results to include in the response."),
+    user_id: Optional[str] = Header(None, alias="user_id"),  # Extract user_id from header, default to None if not present
 ):
     """
     Retrieve the memories in an agent's archival memory store (paginated query).
     """
-    actor = server.get_current_user()
+    actor = server.get_user_or_default(user_id=user_id)
 
     # TODO need to add support for non-postgres here
     # chroma will throw:
@@ -221,11 +227,12 @@ def insert_agent_archival_memory(
     agent_id: str,
     request: CreateArchivalMemory = Body(...),
     server: "SyncServer" = Depends(get_letta_server),
+    user_id: Optional[str] = Header(None, alias="user_id"),  # Extract user_id from header, default to None if not present
 ):
     """
     Insert a memory into an agent's archival memory store.
     """
-    actor = server.get_current_user()
+    actor = server.get_user_or_default(user_id=user_id)
 
     return server.insert_archival_memory(user_id=actor.id, agent_id=agent_id, memory_contents=request.text)
 
@@ -238,11 +245,12 @@ def delete_agent_archival_memory(
     memory_id: str,
     # memory_id: str = Query(..., description="Unique ID of the memory to be deleted."),
     server: "SyncServer" = Depends(get_letta_server),
+    user_id: Optional[str] = Header(None, alias="user_id"),  # Extract user_id from header, default to None if not present
 ):
     """
     Delete a memory from an agent's archival memory store.
     """
-    actor = server.get_current_user()
+    actor = server.get_user_or_default(user_id=user_id)
 
     server.delete_archival_memory(user_id=actor.id, agent_id=agent_id, memory_id=memory_id)
     return JSONResponse(status_code=status.HTTP_200_OK, content={"message": f"Memory id={memory_id} successfully deleted"})
@@ -268,11 +276,12 @@ def get_agent_messages(
         DEFAULT_MESSAGE_TOOL_KWARG,
         description="[Only applicable if use_assistant_message is True] The name of the message argument in the designated message tool.",
     ),
+    user_id: Optional[str] = Header(None, alias="user_id"),  # Extract user_id from header, default to None if not present
 ):
     """
     Retrieve message history for an agent.
     """
-    actor = server.get_current_user()
+    actor = server.get_user_or_default(user_id=user_id)
 
     return server.get_agent_recall_cursor(
         user_id=actor.id,
@@ -306,13 +315,14 @@ async def send_message(
     agent_id: str,
     server: SyncServer = Depends(get_letta_server),
     request: LettaRequest = Body(...),
+    user_id: Optional[str] = Header(None, alias="user_id"),  # Extract user_id from header, default to None if not present
 ):
     """
     Process a user message and return the agent's response.
     This endpoint accepts a message from a user and processes it through the agent.
     It can optionally stream the response if 'stream_steps' or 'stream_tokens' is set to True.
     """
-    actor = server.get_current_user()
+    actor = server.get_user_or_default(user_id=user_id)
 
     # TODO(charles): support sending multiple messages
     assert len(request.messages) == 1, f"Multiple messages not supported: {request.messages}"

letta/server/rest_api/routers/v1/blocks.py

@@ -1,6 +1,6 @@
 from typing import TYPE_CHECKING, List, Optional
 
-from fastapi import APIRouter, Body, Depends, HTTPException, Query
+from fastapi import APIRouter, Body, Depends, Header, HTTPException, Query
 
 from letta.schemas.block import Block, CreateBlock, UpdateBlock
 from letta.server.rest_api.utils import get_letta_server
@@ -19,8 +19,9 @@ def list_blocks(
     templates_only: bool = Query(True, description="Whether to include only templates"),
     name: Optional[str] = Query(None, description="Name of the block"),
     server: SyncServer = Depends(get_letta_server),
+    user_id: Optional[str] = Header(None, alias="user_id"),  # Extract user_id from header, default to None if not present
 ):
-    actor = server.get_current_user()
+    actor = server.get_user_or_default(user_id=user_id)
 
     blocks = server.get_blocks(user_id=actor.id, label=label, template=templates_only, name=name)
     if blocks is None:
@@ -32,8 +33,9 @@ def list_blocks(
 def create_block(
     create_block: CreateBlock = Body(...),
     server: SyncServer = Depends(get_letta_server),
+    user_id: Optional[str] = Header(None, alias="user_id"),  # Extract user_id from header, default to None if not present
 ):
-    actor = server.get_current_user()
+    actor = server.get_user_or_default(user_id=user_id)
 
     create_block.user_id = actor.id
     return server.create_block(user_id=actor.id, request=create_block)