> For the complete documentation index, see [llms.txt](https://docs.supervisely.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.supervisely.com/import-and-export/import/supported-annotation-formats/audio/audio.md).

# Audio recordings

## Overview

This option uploads audio recordings to the platform without any annotations. All recordings from the input directory and its subdirectories are uploaded to a single dataset.

## Format description

**Supported audio formats:** `.wav`, `.flac`, `.mp3`, `.ogg`, `.m4a`\
**With annotations:** No\
**Supported annotation format:** Not applicable.\
**Grouped by:** Any structure (will be uploaded to a single dataset).<br>

## Input files structure

```
📦 folder
┣ 📂 bench_a
┃ ┣ 🎵 run_01.wav
┃ ┗ 🎵 run_02.flac
┗ 📂 bench_b
  ┣ 🎵 run_03.mp3
  ┗ 🎵 run_04.m4a
```

Files with other extensions are skipped. Recordings with the same name get a free name in the dataset.

Audio is always uploaded; it cannot be imported as links.