BerriAI/litellm · error · NotImplementedError

GoogleAIStudioFilesHandler does not support file content…

Error message

GoogleAIStudioFilesHandler does not support file content retrieval

What it means

GoogleAIStudioFilesHandler.transform_file_content_request raises NotImplementedError: downloading raw file content through LiteLLM's file-content interface is not implemented for the Gemini provider. The handler only supports create (two-step resumable upload), retrieve metadata, and delete.

Solutions

Store the original bytes locally (or in your own object storage) at upload time instead of re-downloading from Gemini
For content you own, re-fetch from your source of truth rather than the Gemini Files API
If you must download, check whether Google's Files API supports :download for your file type and call it directly with the x-goog-api-key header
Open an upstream feature request with BerriAI/litellm

Example fix

# before
content = litellm.file_content(file_id='files/abc123', custom_llm_provider='gemini', api_key=KEY)
# NotImplementedError: ...does not support file content retrieval

# after - keep the source bytes; Gemini files are inputs, not storage
def upload_and_keep(path: str):
    data = open(path, 'rb').read()
    obj = litellm.create_file(file=data, custom_llm_provider='gemini', api_key=KEY)
    my_store[obj.id] = data  # serve content from your own store
    return obj

Defensive patterns

Strategy: fallback

Validate before calling

def can_download(provider: str) -> bool:
    return provider != 'gemini'  # content retrieval unsupported by the gemini handler

Prevention

Treat Gemini uploads as one-way inputs: keep the authoritative bytes in your own storage
Branch file-content flows per provider instead of assuming a uniform files API
Add NotImplementedError to the expected-exception list in cross-provider files tests

When it happens

Trigger: Calling litellm.file_content / the files handler's content-download path with custom_llm_provider='gemini' — e.g. trying to fetch bytes of a previously uploaded Gemini file through the OpenAI-style API.

Common situations: Apps that mirror user uploads back to clients via the same abstraction used to upload; debugging workflows that dump file contents; assuming full OpenAI Files API parity for Gemini.

Understand the failure class

Background: UnsupportedOperationException and "is not supported" errors: when a library deliberately refuses a call — this error's family across 30 libraries.

Related errors

AI-assisted analysis of BerriAI/litellm@77b7c6c40c (2026-08-18). Data as JSON: /api/errors/1c22e00b338f9401. Report an issue: GitHub.

Appendix: source

Thrown at litellm/llms/gemini/files/transformation.py:388

        litellm_params: dict,
    ) -> tuple[str, dict]:
        raise NotImplementedError("GoogleAIStudioFilesHandler does not support file listing")

    def transform_list_files_response(
        self,
        raw_response: httpx.Response,
        logging_obj: LiteLLMLoggingObj,
        litellm_params: dict,
    ) -> list[OpenAIFileObject]:
        raise NotImplementedError("GoogleAIStudioFilesHandler does not support file listing")

    def transform_file_content_request(
        self,
        file_content_request,
        optional_params: dict,
        litellm_params: dict,
    ) -> tuple[str, dict]:
        raise NotImplementedError("GoogleAIStudioFilesHandler does not support file content retrieval")

    def transform_file_content_response(
        self,
        raw_response: httpx.Response,
        logging_obj: LiteLLMLoggingObj,
        litellm_params: dict,
    ) -> HttpxBinaryResponseContent:
        raise NotImplementedError("GoogleAIStudioFilesHandler does not support file content retrieval")

View on GitHub (pinned to 77b7c6c40c)