Zackriya-Solutions/meetily · warning

File too large: {:.2}GB. Maximum supported size is {}GB

Error message

File too large: {:.2}GB. Maximum supported size is {}GB

What it means

validate_audio_file (import.rs:149) rejects files whose metadata len() exceeds MAX_FILE_SIZE_BYTES, a 20 GiB cap defined at import.rs:61 explicitly to prevent OOM during the decode-to-memory pipeline (all samples accumulate into a Vec<f32>, multiplying size several-fold over the compressed input) and to bound processing time. It fires after existence/format checks and before duration probing. The cap is on total file size, not audio payload, so video-heavy MP4s hit it even though their audio track is small.

Source

Thrown at frontend/src-tauri/src/audio/import.rs:149

        .map(|e| e.to_lowercase())
        .unwrap_or_default();

    if !AUDIO_EXTENSIONS.contains(&extension.as_str()) {
        return Err(anyhow!(
            "Unsupported format: .{}. Supported: {}",
            extension,
            AUDIO_EXTENSIONS.join(", ")
        ));
    }

    // Get file size
    let metadata = std::fs::metadata(path)
        .map_err(|e| anyhow!("Cannot read file: {}", e))?;
    let size_bytes = metadata.len();

    // Check file size limit
    if size_bytes > MAX_FILE_SIZE_BYTES {
        return Err(anyhow!(
            "File too large: {:.2}GB. Maximum supported size is {}GB",
            size_bytes as f64 / (1024.0 * 1024.0 * 1024.0),
            MAX_FILE_SIZE_BYTES / (1024 * 1024 * 1024)
        ));
    }

    // Get filename without extension for title
    let filename = path
        .file_stem()
        .and_then(|s| s.to_str())
        .unwrap_or("Imported Audio")
        .to_string();

    // Try fast metadata-only validation first
    let duration_seconds = match extract_duration_from_metadata(path) {
        Ok(duration) => {
            debug!(
                "Got duration from metadata: {:.2}s (fast path)",

View on GitHub (pinned to 0281737d87)

Solutions

1. Extract just the audio: ffmpeg -i huge.mp4 -vn -c:a aac -b:a 64k audio.m4a - usually shrinks 20+ GiB video to a few hundred MB.
2. Split long recordings into parts under 20 GiB (ffmpeg -i in.wav -f segment -segment_time 3600 part%02d.wav) and import sequentially.
3. Downsample while extracting (-ac 1 -ar 16000) to match the Whisper pipeline target and cut resample memory too.
4. If >20 GiB single-file import is truly needed, raise MAX_FILE_SIZE_BYTES in import.rs - but first verify RAM headroom for the in-memory f32 samples.
5. Prevent at the UI: show the size limit next to the picker so users transcode before selecting.

Example fix

# before: 25GB mp4 rejected by the 20GiB cap
# after: strip video, keep speech-quality audio - imports fine
ffmpeg -i meeting-4k.mp4 -vn -ac 1 -ar 16000 -c:a aac -b:a 48k meeting-audio.m4a

// Frontend: pre-check size before invoking so users get guidance, not an error
const MAX = 20 * 1024 * 1024 * 1024;
const stat = await statFile(path); // via fs plugin
if (stat.size > MAX) suggestAudioExtraction(path);

const withinImportLimit = (sizeBytes: number) => sizeBytes <= 20 * 1024 * 1024 * 1024;

try { await invoke('import_audio_file', payload); }
catch (e) {
  if (String(e).startsWith('File too large'))
    showHint('Run: ffmpeg -i file.mp4 -vn -ac 1 -ar 16000 -c:a aac out.m4a and import out.m4a');
  else throw e;
}

When it happens

Trigger: Selecting multi-hour lossless captures (24h WAV at 48kHz stereo is ~15.5 GiB uncompressed - borderline); huge video files whose audio is tiny but whose container exceeds 20 GiB (4K screen recordings, multi-hour seminar MP4s); concatenated archives; PMR/recovery recordings.

Common situations: Importing day-long conference recordings or screen captures; WAV masters from studio sessions; users assuming audio import counts only the audio stream; surveillance/lecture capture archives.

Related errors

• File does not exist: {}
• Unsupported format: .{}. Supported: {}
• Cannot read file: {}
• Source file not found: {}
• No transcript text available. Please add some text first.

AI-assisted analysis of Zackriya-Solutions/meetily@0281737d87 (2026-08-16). Data as JSON: /api/errors/ccdbe8480ca6b407. Report an issue: GitHub.