How to Use StemNook: Vocals, Backing Tracks and WAV
Learn how to use StemNook to separate vocals and backing tracks, check your free minutes, listen to both outputs, download WAV files and handle common upload issues.
The short answer: Choose an MP3 or WAV, click Start, sign in with Google if needed, then return and start separation. Listen to Instrumental for the backing track or Vocals for the voice, then download the WAV you want. Free accounts receive 600 seconds of processing time per UTC calendar month.
You have a song and want to sing along without competing with the original singer. Or you want to hear the vocal phrasing more clearly. Those tasks start with the same file, but you will choose a different output at the end.
I would start by deciding which track I need before opening the tool. That keeps the next steps straightforward: prepare the file, check the account, process it once and listen before saving.
This guide follows the shared separation workflow on the homepage, Vocal Remover and Acapella Extractor pages. The inner pages prioritize Instrumental or Vocals respectively. Check the tool’s availability before submitting.
The pictures below distinguish captured interface states from a result-image placeholder. A picture of a selected file is not proof that a cloud job finished, and the result slot does not claim a measured separation outcome.
1. Choose the result you actually need
StemNook separates a recording into two tracks. It does not take a finished song apart into every original studio recording.
| Your task | Track to use | What to check |
|---|---|---|
| Sing or play along | Instrumental | Voice left behind, missing instruments and rough sounds |
| Study the singing | Vocals | Instrument bleed, echoes and the start and end of phrases |
| Cut out one passage | Browser audio cutter | Start and end times, then the downloaded clip |
Instrumental means the combined accompaniment. Vocals means the extracted voice. A separate drum, bass or guitar file needs another workflow; these are not outputs of this two-track tool.
For karaoke, the Instrumental file gives you audio. Lyrics on screen, synchronized words and CDG files are separate tasks. Getting the accompaniment ready is a useful first step, but it does not automatically create a complete karaoke video.
2. Choose a file, then sign in when you start
Open StemNook’s separation workspace and choose your song. Processing requires Google sign-in; the Start action takes you there if needed. You can also sign in before choosing a file.
The free allowance is 10 minutes, or 600 seconds, per UTC calendar month. Check the allowance shown for your account rather than assuming that every visit begins with a new ten minutes. This is a monthly processing allowance, not an unlimited-free offer.
A file can also need more time than you have left without the balance being zero. For example, if your account shows less time than the recording’s duration, choose a shorter eligible file or check the available account options.
The browser tries to keep your selected file for up to 30 minutes during sign-in or checkout. On return, check the filename and start when ready; nothing is submitted automatically. If storage is unavailable or expired, select your file again.
The current local screenshots do not show a completed Google sign-in or a live account balance. This is an account step you need to complete in your own session; no screenshots contain another person’s account information.
3. Prepare an MP3 or WAV that fits the limits
Before choosing a file, check three things:
- The file is an MP3 or WAV you have on your device.
- It is no larger than the interface’s 30 MB limit.
- Its duration is between 0.5 seconds and 5 minutes.
Both size and duration apply. A song can be shorter than five minutes and still be too large, particularly if it is a WAV.
For perspective, 44.1 kHz, 16-bit stereo PCM WAV takes about 10.6 MB per minute. That is a calculation from the format, not a measurement of your file. The actual 30 MiB byte cap means that this kind of input WAV reaches the size limit just before three minutes.
If your recording does not fit, prepare a shorter authorized excerpt in an audio editor. Do not just rename another format to .mp3 or .wav: changing a file name does not change the audio encoding.
Choose the prepared file and check its displayed name. The earlier screenshot below shows a selected file with processing unavailable. In the current workflow, Start prompts Google sign-in when needed. It is a file-preparation example, not a finished cloud separation.
4. Start one separation and keep track of its status
Once you are signed in and the file is valid, use Separate vocals & music. Let the existing task finish before creating another one.
The input duration is rounded up to whole seconds for the task’s quota reservation. That means a recording with a fraction of a second is not treated as a shorter whole-second file.
If the page cannot retrieve progress, check the task’s status again before submitting another job. A progress-query problem and a confirmed processing failure are different situations. Repeatedly clicking the action is not a reliable way to recover either of them.
There is release logic for known failed tasks, while an uncertain submission can keep its reservation until its status is reconciled. This guide does not promise that every failure or retry is free.
The next image position is reserved for a real completed task. Until that verification is available, it is labelled as a placeholder.
Step diagram · screenshot pending
A real result will go here
Vocals: check the voice
Instrumental: check the backing
Download the track you need
5. Listen to the track that serves your task
For singing practice, start with Instrumental. For vocal study, start with Vocals. Listen before downloading both large files.
Keep the original in your usual audio player and compare the same moments in StemNook. A simple check is enough:
- Pick a verse, a chorus and the ending in your original.
- Listen to those moments in the output you want.
- Check whether important words or instruments disappear.
- Listen for leftover voice, echoes or metallic sounds.
- Decide whether the result is useful for your particular practice or edit.
For a short clip, listen from beginning to end. You do not need to invent three checkpoints inside a seven-second recording.
The built-in listening example can help you find the players, but it is a short synthetic spoken example, not a singing-quality test. It cannot tell you how your own song will sound after processing.
AI separation can leave voice in the accompaniment or instruments in the vocal track. It may also retain reverb. Neither output should be presented as the original dry studio recording or as a promise of completely clean audio.
6. Download WAV while your results are available
The separation results are 44.1 kHz, 16-bit PCM stereo WAV. StemNook does not directly export these results as MP3.
Download the track you need within 24 hours after completion on a free account. That is the result-access window; it does not establish a broader file-deletion policy.
WAV files are larger than many compressed inputs. Two four-minute outputs at this format total about 84.7 MB, based on the format calculation. This is not a measured download size or a promise about every file.
If you later want one short passage, you can open an eligible file in the browser audio cutter. That is a separate local workflow. There is no one-click transfer promised here, and a long exported WAV may exceed the cutter’s input-size limit.
7. Fix the problem you are actually seeing
The file will not open
Check the extension, size and duration first. Try another valid MP3 or WAV. If the recording is damaged or the browser cannot decode it, re-export it from the original source.
The separation action is unavailable
Check whether you are signed in, whether the file is selected and whether the service is available. Selecting a file does not, by itself, create an account or start processing.
There is not enough time left this month
Check the remaining allowance against this file. “Not enough for this file” is different from “nothing left.” Choose an eligible shorter recording or review your account options.
Progress cannot be retrieved
Check the existing task first. A network interruption does not prove that the task stopped, and starting again may create another task.
A result link has expired
Free results are available for 24 hours; V4 Pro and Studio results are available for 7 and 30 days respectively. Check the actual download deadline shown with your task. Keep your source file and save usable outputs while the links are available.
Start with one recording and one clear task
Choose a short recording you are allowed to edit. Decide whether you want the voice or the accompaniment, then follow the homepage workflow and listen before saving.
For more detail, read the vocal remover tool or the acapella extractor tool. These explain different uses of the same two outputs; they do not add extra instrument stems.