xtekky/gpt4free · error · ValueError

Invalid input audio

Error message

Invalid input audio

What it means

Thrown by g4f.image.to_input_audio() when audio is passed as a non-str object (bytes, Path, file-like) together with a filename, but get_extension(filename) returns None — i.e. the filename has no dot or its extension is not in EXTENSIONS_MAP. The library needs a known audio extension to fill the 'format' field of the input-audio dict, so it refuses the call.

Source

Thrown at g4f/image/__init__.py:514

        except (AttributeError, io.UnsupportedOperation):
            pass
        return image.read()


def to_data_uri(image: ImageType, filename: str = None) -> str:
    if not isinstance(image, str):
        data = to_bytes(image)
        data_base64 = base64.b64encode(data).decode()
        return f"data:{is_data_an_media(data, filename)};base64,{data_base64}"
    return image


def to_input_audio(audio: ImageType, filename: str = None) -> str:
    if not isinstance(audio, str):
        if filename is not None:
            format = get_extension(filename)
            if format is None:
                raise ValueError("Invalid input audio")
            return {
                "data": base64.b64encode(to_bytes(audio)).decode(),
                "format": format,
            }
        raise ValueError("Invalid input audio")
    audio = re.match(r"^data:audio/(\w+);base64,(.+?)", audio)
    if audio:
        return {"data": audio.group(2), "format": audio.group(1).replace("mpeg", "mp3")}
    raise ValueError("Invalid input audio")


def use_aspect_ratio(extra_body: dict, aspect_ratio: str) -> Image:
    extra_body = {key: value for key, value in extra_body.items() if value is not None}
    if extra_body.get("width") is None or extra_body.get("height") is None:
        width, height = get_width_height(
            aspect_ratio, extra_body.get("width"), extra_body.get("height")
        )
        extra_body = {"width": width, "height": height, **extra_body}

View on GitHub (pinned to 973504e177)

Solutions

1. Give the filename a recognized audio extension: use .mp3, .wav, .ogg, .webm, .flac, .aac, or another key present in g4f.image.EXTENSIONS_MAP.
2. Check beforehand: from g4f.image import get_extension; assert get_extension(filename) is not None.
3. If the format genuinely is unsupported, convert the audio to mp3/wav (e.g. with pydub) before calling.

Example fix

// before
payload = to_input_audio(open('voicenote','rb').read(), filename='voicenote')

// after
payload = to_input_audio(open('voicenote.mp3','rb').read(), filename='voicenote.mp3')

from g4f.image import get_extension

def valid_audio_filename(filename: str) -> bool:
    return get_extension(filename) is not None

assert valid_audio_filename("clip.mp3")

When it happens

Trigger: Calling to_input_audio(audio_bytes, filename="recording") with no extension, or filename="track.opus" / "audio.m4a" where the extension is not registered in g4f's EXTENSIONS_MAP.

Common situations: Generating temp filenames like tempfile.mkstemp() suffix-less names; using rarer audio containers the map does not cover; upstream code deriving the filename from a URL query string that lost its extension.

Related errors

• No input provided
• Invalid image format or file not found. Expected bytes, str,
• Response must be a dict or have headers
• url must be a non-empty string
• url must start with http:// or https://

AI-assisted analysis of xtekky/gpt4free@973504e177 (2026-08-14). Data as JSON: /api/errors/59cce051b01e4276. Report an issue: GitHub.