> ## Documentation Index
> Fetch the complete documentation index at: https://docs.audiopod.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Remove background noise from audio

> Clean hiss, hum, wind, traffic and room noise out of voice recordings and videos with the AudioPod Python or Node SDK, with three quality modes.

**What you get:** a cleaned copy of your recording as a download URL. For a video input you also get the video back with the cleaned audio.

**Input:** a local audio or video file, or a URL of one.

## 1. Install and set your key

<CodeGroup>
  ```bash Python theme={null}
  pip install audiopod
  export AUDIOPOD_API_KEY="ap_..."
  ```

  ```bash Node theme={null}
  npm install audiopod
  export AUDIOPOD_API_KEY="ap_..."
  ```
</CodeGroup>

## 2. Denoise

<CodeGroup>
  ```python Python theme={null}
  from audiopod import Client

  client = Client()

  # denoise() waits for the job by default
  job = client.denoiser.denoise(file="voice-memo.m4a")
  print(job["output_url"])
  ```

  ```typescript Node theme={null}
  import AudioPod from "audiopod";

  const client = new AudioPod();

  // denoise() submits the job and waits for it
  const job = await client.denoiser.denoise({ file: "./voice-memo.m4a" });
  console.log(job.output_url);
  ```
</CodeGroup>

For a video, read `video_output_url` for the cleaned video as well as `output_url` for the audio.

## Quality modes

| Mode | Use it for |
| - | - |
| `balanced` | The fastest mode, for quick clean-ups and previews. |
| `studio` (default) | Broadcast and podcast quality. |
| `ultra` | The highest fidelity, restoring lost high frequencies. |

Each mode is priced per minute of audio; see [pricing](https://audiopod.ai/pricing). In the REST API the field is `quality_mode`. The SDKs take it as `mode` from their next release; until then, leave it out to get `studio`.

## Tips

* Clean voice samples before [cloning a voice](/quickstarts/voice-cloning) and noisy recordings before [transcription](/quickstarts/transcription): both come out better.
* Music in the background is not "noise". To pull a voice out of a song, use [stem separation](/quickstarts/stem-separation) in `two` mode instead.

## From an AI agent

With the [AudioPod MCP server](/agents) connected (core or full): *"Clean up the background noise in [https://example.com/voice-memo.m4a](https://example.com/voice-memo.m4a)."* The agent calls `denoise_audio` with `file_url`.

Full endpoint reference: [Noise Reduction](/api-reference/noise-reduction).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.