agentscope-ai/agentscope · error · ValueError

"AgentScopeLLM requires `model` in the config to be an Agent

Error message

"AgentScopeLLM requires `model` in the config to be an AgentScope ChatModelBase instance."

What it means

AgentScopeLLM is a mem0 LLM adapter that requires an AgentScope ChatModelBase instance to be supplied via config['model']. The constructor checks that model is set before validating its type; None raises this ValueError. Without it there is no underlying chat model for mem0 to route calls through.

Source

Thrown at src/agentscope/middleware/_longterm_memory/_mem0/_agentscope_adapter.py:110


class AgentScopeLLM(LLMBase):
    """mem0 ``LLMBase`` backed by an AgentScope ``ChatModelBase``.

    Pass your AgentScope model into ``config["model"]``; mem0's memory
    extraction calls then route through it. Both streaming and
    non-streaming AgentScope models are accepted (streaming responses are
    drained and the final chunk is used).
    """

    def __init__(
        self,
        config: BaseLlmConfig | dict | None = None,
    ) -> None:
        """Initialize the AgentScope LLM for mem0."""
        super().__init__(config)
        if self.config.model is None:
            raise ValueError(
                "AgentScopeLLM requires `model` in the config to be an "
                "AgentScope ChatModelBase instance.",
            )
        if not isinstance(self.config.model, ChatModelBase):
            raise TypeError(
                f"AgentScopeLLM `model` must be a ChatModelBase, got "
                f"{type(self.config.model).__name__}.",
            )
        self._agentscope_model: ChatModelBase = self.config.model
        self._bridge = _AsyncBridge()

    # ----- LLMBase interface -----
    # pylint: disable=unused-argument
    def generate_response(
        self,
        messages: list[dict[str, str]],
        response_format: Any | None = None,  # mem0 contract — unused
        tools: list[dict] | None = None,

View on GitHub (pinned to e90f1c7592)

Solutions

Pass an AgentScope ChatModelBase instance (e.g. OpenAIChat(model='gpt-4o-mini')) in the config's model field
If using Mem0Middleware, pass chat_model=... instead of hand-building the LLM config
Use build_mem0_config(chat_model=..., embedding_model=...) which constructs the config correctly

Example fix

// before
llm = AgentScopeLLM(BaseLlmConfig(provider='agentscope'))
// after
from agentscope.model import OpenAIChat
llm = AgentScopeLLM(BaseLlmConfig(model=OpenAIChat(model='gpt-4o-mini')))

Defensive patterns

Strategy: validation

Validate before calling

from agentscope.model import ChatModelBase
cfg = my_config
if getattr(cfg, 'model', None) is None:
    raise ValueError('config.model must be a ChatModelBase instance')

Type guard

from agentscope.model import ChatModelBase
def has_chat_model(cfg) -> bool:
    return isinstance(getattr(cfg, 'model', None), ChatModelBase)

Try / catch

try:
    llm = AgentScopeLLM(cfg)
except ValueError as e:
    if 'model' in str(e):
        cfg['model'] = OpenAIChat(model='gpt-4o-mini')
        llm = AgentScopeLLM(cfg)
    else:
        raise

Prevention

Always construct adapters via build_mem0_config rather than by hand
Assert config.model is an AgentScope model object before constructing AgentScopeLLM

When it happens

Trigger: Constructing AgentScopeLLM(BaseLlmConfig(...)) or AgentScopeLLM({'model': None, ...}) without setting the model field; e.g. copying a mem0 OpenAI config dict and omitting/replacing 'model' with None.

Common situations: Porting an existing mem0 LLM config to agentscope; passing provider/api_key-style dicts where a model string was expected instead of a ChatModelBase object; building Mem0Middleware config manually.

Related errors

AI-assisted analysis of agentscope-ai/agentscope@e90f1c7592 (2026-08-28). Data as JSON: /api/errors/90afa65d3dea8e1c. Report an issue: GitHub.