agentscope-ai/agentscope · error · StructuredOutputError

Failed to get the completed response from model {model_name}

Error message

Failed to get the completed response from model {model_name}.

What it means

After invoking the model with a structured-output strategy, the completed response was None or had empty content, so agentscope raises StructuredOutputError. The API call technically returned but yielded no usable text (e.g. empty choices/content, streaming accumulation failure, or content filtered).

Source

Thrown at src/agentscope/model/_base.py:693

            # avoid duplicating the retry logic in ``__call__``), we must
            # replicate that accumulation here, otherwise the stream may
            # end without ever producing an ``is_last=True`` chunk.
            acc_res = _StreamAccumulator()

            async for chunk in res:
                if chunk.is_last:
                    completed_response = chunk
                    break
                acc_res.append_chat_response(chunk)
                acc_res.id = chunk.id

            if completed_response is None:
                completed_response = acc_res.build()
        else:
            completed_response = res

        if completed_response is None or not completed_response.content:
            raise StructuredOutputError(
                f"Failed to get the completed response from model "
                f"{model_name}.",
            )

        structured_output: dict[str, Any] | None = None
        try:
            for _ in completed_response.content:
                if isinstance(_, ToolCallBlock) and _.name == func_name:
                    structured_output = _json_loads_with_repair(
                        _.input,
                        input_schema,
                    )
                    break

            if structured_output is None:
                raise StructuredOutputError(
                    "Failed to generate structured output for model.",
                )

View on GitHub (pinned to e90f1c7592)

Solutions

1. Retry the call — transient empty responses from providers are common
2. Increase max_tokens and simplify/strengthen the prompt so the model actually emits the JSON
3. Inspect the raw provider response (enable verbose logging) to see finish_reason/empty content cause
4. If content filtering is the cause, adjust the prompt or safety settings

Example fix

# before
res = await model.generate_structured_output(msgs, Schema)  # empty content

# after
for attempt in range(3):
    try:
        res = await model.generate_structured_output(msgs, Schema)
        break
    except StructuredOutputError:
        if attempt == 2:
            raise

from agentscope.model import StructuredOutputError
for attempt in range(3):
    try:
        return await model.generate_structured_output(msgs, Schema)
    except StructuredOutputError:
        if attempt == 2:
            raise
        await asyncio.sleep(2 ** attempt)

When it happens

Trigger: Model returns an empty completion (empty content string), stream finished without accumulating any content, or provider returns finish_reason content_filter with no text.

Common situations: Safety filter blocking the output; max_tokens set so low the model emits nothing; provider outage returning empty bodies; prompt asking model to return nothing.

Related errors

• Model call failed after retries, but no exception was raised
• "AgentScope streaming model yielded no chunks."
• "AgentScope embedding model returned no embeddings."
• The input messages cannot be empty for the `generate_structu
• No structured-output strategy is available for {self.model}.

AI-assisted analysis of agentscope-ai/agentscope@e90f1c7592 (2026-08-28). Data as JSON: /api/errors/8d32a5144a08ac17. Report an issue: GitHub.