BerriAI/litellm · error · ValueError

Unable to infer MIME type from GCS URL: {gcs_url}. Supported

Error message

Unable to infer MIME type from GCS URL: {gcs_url}. Supported extensions: {', '.join(extension_to_mime.keys())}

What it means

Gemini embedding MIME inference guard: the GCS URL's extension is not in the supported set (.png/.jpg/.jpeg/.mp3/.wav/.mp4/.mov/.pdf), so no mime_type can be inferred for the fileData part and the URL is rejected.

Source

Thrown at litellm/llms/vertex_ai/gemini_embeddings/batch_embed_content_transformation.py:82

        ValueError: If file extension is not supported
    """
    extension_to_mime: Final = {
        ".png": "image/png",
        ".jpg": "image/jpeg",
        ".jpeg": "image/jpeg",
        ".mp3": "audio/mpeg",
        ".wav": "audio/wav",
        ".mp4": "video/mp4",
        ".mov": "video/quicktime",
        ".pdf": "application/pdf",
    }

    gcs_url_lower: Final = gcs_url.lower()
    for ext, mime_type in extension_to_mime.items():
        if gcs_url_lower.endswith(ext):
            return mime_type

    raise ValueError(
        f"Unable to infer MIME type from GCS URL: {gcs_url}. "
        f"Supported extensions: {', '.join(extension_to_mime.keys())}"
    )


def _parse_data_url(data_url: str) -> tuple[str, str]:
    """
    Parse a data URL to extract the media type and base64 data.

    Args:
        data_url: Data URL in format: data:image/jpeg;base64,/9j/4AAQ...

    Returns:
        tuple: (media_type, base64_data)
            media_type: e.g., "image/jpeg", "video/mp4", "audio/mpeg"
            base64_data: The base64-encoded data without the prefix

    Raises:

View on GitHub (pinned to 77b7c6c40c)

Solutions

1. Use a GCS URL whose extension maps to a supported MIME type (see the listed extensions).
2. Rename or convert the file to a supported type.

Example fix

# use e.g. .pdf, .png, .jpg files in the GCS URL.

When it happens

Trigger: Triggered when the MIME type cannot be inferred from the GCS URL extension for Gemini batch embedding.

Common situations: See trigger scenarios.

AI-assisted analysis of BerriAI/litellm@77b7c6c40c (2026-08-18). Data as JSON: /api/errors/57147c7d7f219584. Report an issue: GitHub.