openai/codex · error

unsupported audio format

Error message

unsupported audio format

What it means

io::Error(ErrorKind::InvalidData, 'unsupported audio format') from snapshot_local_user_input (codex-rs/protocol/src/local_media.rs:37-39): a UserInput::LocalAudio path failed audio_mime_for_path (local_media.rs:74-89), which maps only .wav, .mp3, .m4a, .webm, and .ogg (case-insensitive) to MIME types. Detection is purely extension-based - content is never sniffed - and files with any other or missing extension are rejected before being read.

Source

Thrown at codex-rs/protocol/src/local_media.rs:38

    match input {
        UserInput::LocalImage { path, detail } => {
            let image_detail = *detail;
            let mode = match image_detail {
                Some(ImageDetail::Original) => PromptImageMode::Original,
                Some(ImageDetail::Auto | ImageDetail::Low | ImageDetail::High) | None => {
                    PromptImageMode::ResizeToFit
                }
            };
            let file_bytes = read_bounded_local_media(path, MAX_PROMPT_IMAGE_INPUT_BYTES, "image")?;
            let image = load_for_prompt_bytes(path, file_bytes, mode).map_err(io::Error::other)?;
            *input = UserInput::Image {
                image_url: image.into_data_url(),
                detail: image_detail,
            };
        }
        UserInput::LocalAudio { path } => {
            let mime = audio_mime_for_path(path).ok_or_else(|| {
                io::Error::new(io::ErrorKind::InvalidData, "unsupported audio format")
            })?;
            let file_bytes = read_bounded_local_media(path, MAX_PROMPT_AUDIO_INPUT_BYTES, "audio")?;
            *input = UserInput::Audio {
                audio_url: data_url_from_bytes(mime, &file_bytes),
            };
        }
        UserInput::Text { .. }
        | UserInput::Image { .. }
        | UserInput::Audio { .. }
        | UserInput::Skill { .. }
        | UserInput::Mention { .. } => {}
    }
    Ok(())
}

fn read_bounded_local_media(path: &Path, max_bytes: usize, kind: &str) -> io::Result<Vec<u8>> {
    let file = std::fs::File::open(path)?;
    if file.metadata()?.len() > max_bytes as u64 {

View on GitHub (pinned to 339751715c)

Solutions

1. Convert the audio to a supported container (WAV, MP3, M4A, WebM, or OGG), e.g. ffmpeg -i in.flac out.mp3.
2. If the content is already a supported format but misnamed, rename it so the extension matches the actual container.
3. Verify the final extension before queueing - matching is extension-only, so the name must agree with the bytes.

Example fix

// before: attaching an unsupported file
UserInput::LocalAudio { path: Path::new("/tmp/interview.flac").into() }
// snapshot_local_user_input -> InvalidData: unsupported audio format

// after: convert to a supported format first
// ffmpeg -i interview.flac interview.mp3
UserInput::LocalAudio { path: Path::new("/tmp/interview.mp3").into() }

const SUPPORTED: &[&str] = &["wav", "mp3", "m4a", "webm", "ogg"];
fn audio_supported(path: &Path) -> bool {
    path.extension()
        .and_then(|e| e.to_str())
        .is_some_and(|e| SUPPORTED.iter().any(|s| e.eq_ignore_ascii_case(s)))
}
// gate before snapshotting:
if !audio_supported(&path) {
    reject_with_hint(&path);
}

fn is_unsupported_audio(err: &io::Error) -> bool {
    err.kind() == io::ErrorKind::InvalidData
        && err.to_string().contains("unsupported audio format")
}

match snapshot_local_user_input(&mut input) {
    Err(e) if e.kind() == io::ErrorKind::InvalidData && e.to_string().contains("unsupported audio format") => {
        // tell the user which formats are accepted (wav/mp3/m4a/webm/ogg)
    }
    other => other?,
}

When it happens

Trigger: Queueing UserInput::LocalAudio { path } (via prepare_queued_user_input / snapshot_local_user_input) whose extension is outside wav/mp3/m4a/webm/ogg - for example .flac, .aac, .opus, .txt, or no extension at all.

Common situations: Users attach FLAC or AAC recordings; downloads arrive with unusual extensions; files saved without extensions; double extensions like .mp3.txt.

Related errors

• {kind} input exceeds {max_bytes} bytes
• filename '{filename}' {reason}
• ad-hoc note must not be empty
• path '{path}' {reason}
• line_offset must be a 1-indexed line number

AI-assisted analysis of openai/codex@339751715c (2026-08-25). Data as JSON: /api/errors/0e1228e89d6e6cee. Report an issue: GitHub.