简体中文
EnglishWebsite review
This review may include AI-generated content and is not individually verified. Domain, certificate and other facts come from the observed data on this page.
What is Ezstems?
Ezstems is a free web tool for splitting a song or other audio file into separate "stems" — individual tracks such as vocals, drums, bass and other instruments — using the Spleeter AI model. You upload a file in your browser, wait while it processes, and the resulting stems download automatically. No software installation or audio-editing skills are needed.
What you can control
The page evidence shows the upload form lets you choose:
- Number of stems: 2, 4 or 5
- Output format and bitrate: mp3 at 128, 192, 256 or 320 kbps, or wav
- Frequency cutoff: 11 kHz or 16 kHz
More stems mean finer separation (for example, isolating bass or piano separately rather than lumping them into an "other" track), but also more files to manage. Higher bitrates and wav output preserve more audio detail at the cost of larger downloads.
Who it suits
- DJs and remixers who want an a cappella or instrumental version of a track.
- Producers and students studying how a mix is arranged, or building sample material.
- Singers and cover artists who need a backing track without lead vocals.
- Developers who want separation inside their own apps — the site mentions API access through a paid subscription.
The free-tier trade-offs
The site states free use is capped at 20 uploads per day and a 20 MB file size limit, and that processing may involve a queue unless you subscribe. That makes it fine for occasional tracks, less so for batch work or long, high-quality files. A subscriber tier is offered to bypass those limits.
If you want to judge whether it fits your workflow, test it with one short, familiar song: compare the vocal stem against the original, and check whether the chosen stem count and bitrate give you usable results before committing to a larger project. For more ambitious editing afterward, a full digital audio workstation such as Ardour or Audacity can handle the exported stems.
How do I upload an audio file and get stems?
Upload your track on the Ezstems home page, choose how many stems and which output format you want, then start processing. Ezstems runs your file through the Spleeter source-separation model and the separated parts download automatically when the job finishes. The page shows progress as “Uploading your file…”, then “Processing your song…”, then “Processing Complete!”, so you can tell where you are in the sequence.
The practical steps
- Pick your stem count: 2, 4 or 5 stems.
- Pick an output format and quality from the options shown, such as MP3 at 128, 192, 256 or 320, or WAV with an 11kHz or 16kHz cutoff.
- Accept the terms and conditions.
- Upload the audio file.
- Wait while the page reports upload and processing progress. The page warns you not to close your browser during processing.
- When processing completes, the stems download automatically, with a manual download link as a fallback.
- Use “Process Another” to run a second file without leaving the page.
What to expect from the results
Spleeter separates a mixed song into parts, typically vocals and accompaniment at the simplest setting, with more parts available at higher stem counts. The output is useful for practice and production tasks: pulling a vocal out to learn a melody, removing a vocal for a karaoke-style backing, or isolating a part to study an arrangement. Separation is rarely perfect, so expect some bleed between stems, especially on dense or heavily compressed mixes. WAV output is the better choice if you plan to edit further; MP3 is lighter and faster to handle.
Limits and when to subscribe
The free tier has practical ceilings: a 20 uploads per day limit and a 20MB file size cap. A queue is also shown while processing, and the page offers a way to skip it. If you regularly hit those limits or need programmatic access, Ezstems points to a paid subscription that raises the upload and file-size caps and provides API access. Check Ezstems for the current terms, since limits and formats can change.
A quick decision guide
- One-off vocal or backing track from a single song: the free upload path is enough.
- Several songs per session, or files longer than the size cap allows: a subscription is the practical route.
- Building separation into your own app or workflow: the API access mentioned on the subscription page is the relevant option.
If you only need occasional separation, start with a 2-stem run to check quality on your material before spending time on higher stem counts.
What are the file size and daily upload limits for free users?
Free Ezstems users can upload files up to 20MB each and process up to 20 uploads per day. Both limits are described on the site as applying to the free tier, with the paid subscription offered as the way to bypass them.
What each limit means in practice
- 20MB per file: This is the practical constraint for most users. A 20MB ceiling is generous for a single song in a compressed format such as MP3, but it can be tight for long recordings, high-bitrate files, or lossless formats like WAV. If your file is rejected, converting it to a lower-bitrate MP3 first is the usual workaround — though you should weigh whether the quality loss matters for your intended use.
- 20 uploads per day: This suits occasional use — a handful of songs, a few practice mixes, or experimenting with different stem counts. It becomes restrictive if you are batch-processing a large library or running repeated tests on the same track with different settings.
Who these limits fit
| Use case | Free tier fit |
|---|---|
| Separating a few songs for practice or covers | Comfortable |
| Trying out 2-, 4-, or 5-stem output on a short track | Comfortable |
| Processing a full album or large back catalogue | Likely to hit the daily cap |
| Working with long or lossless source files | May hit the 20MB ceiling |
| Building stem separation into your own app | Not viable on free access |
A useful next step
Before uploading, check your file's size and format. If it exceeds 20MB, decide whether a compressed conversion is acceptable for your purpose. If you expect to process many files or need larger uploads and API access, compare that workload against the subscription option on the site rather than assuming the free tier will stretch. For a rough sense of scale, a typical three-to-four-minute MP3 song usually sits well under 20MB, so the per-file cap mainly affects longer or higher-quality sources.
What features do I get with a paid subscription?
A paid subscription on Ezstems is mainly about removing the free tier's limits rather than unlocking a different separation engine. The site states that subscribing lets you bypass the 20 uploads per day limit, increase the 20MB file size cap, and get direct API access for your own applications. The free page also shows the standard processing options — 2, 4 or 5 stems, output as MP3 at 128/192/256/320 kbps or WAV, and 11kHz or 16kHz cutoff — so check the subscription page itself to confirm whether any of those choices change on a paid plan.
Free vs. subscribed, based on what the site advertises
| Aspect | Free use | Paid subscription |
|---|---|---|
| Uploads | 20 per day | Bypass the daily limit |
| File size | 20MB cap | Higher cap |
| API access | Not mentioned | Direct access for your own apps |
| Stem and format options | 2/4/5 stems, MP3 or WAV, cutoff choice | Confirm on the subscription page |
| Queue | You are placed in a queue | Membership options are offered as a way to skip it |
The queue point is worth noting: while processing, the page tells you not to close your browser and shows your position, with an option to skip the queue via membership. If you are batch-processing a rehearsal recording or a long DJ set, that waiting time plus the 20MB cap is the practical reason to pay; for the occasional single track, the free tier is usually enough.
A concrete way to decide
Suppose you want to practise guitar over a song: 2 stems (vocals and accompaniment) at MP3 320 is plenty, and one upload fits inside the free allowance. Suppose instead you are building a small app that splits uploaded audio automatically — that is exactly the case the API access is aimed at, and the daily limit would otherwise block you.
Next step: open Ezstems and compare the free upload form against the subscription page, checking three things — the new file size cap, whether the stem count and output format options expand, and what API access actually includes (rate limits, authentication, output format). If you only need a handful of splits a week, the free tier's 2/4/5-stem options and MP3/WAV output already cover most home practice and remix work.
What output formats and stem options are available?
Ezstems lets you choose how many stems to split a track into and what audio format to download them in. According to the site's own interface, the core options are:
- Stem counts: 2 stems, 4 stems, or 5 stems
- MP3 output: 128, 192, 256, or 320 kbps
- WAV output: 11 kHz or 16 kHz cutoff
A "2 stems" split is the classic vocal/instrumental separation, which is the most useful choice if you want an acapella or a karaoke-style backing track. "4 stems" typically separates vocals, drums, bass, and "other," and "5 stems" usually adds piano to that set — the standard Spleeter configurations. Ezstems states it processes uploads with the Spleeter model, so the stem counts line up with that tool's known modes.
Choosing format and stem count
| Goal | Sensible choice |
|---|---|
| Karaoke or acapella | 2 stems, MP3 320 or WAV |
| Remixing or practice with individual parts | 4 or 5 stems |
| Quick preview, small files | MP3 128 |
| Further editing or mastering | WAV, higher cutoff |
The trade-off is straightforward: more stems give finer control but more files to manage and more chance of bleed or artifacts in each one. Higher bitrates and the 16 kHz cutoff preserve more high-frequency detail, which matters if you plan to re-edit or re-master; the lower settings are fine for casual listening.
If you want a second opinion on the same task, Moises and Lalal.ai are established alternatives with their own stem and format options, useful for comparing results on a track Spleeter handles poorly.
Practical next step: run one song at 2 stems and the same song at 4 or 5 stems, then listen to the vocal track in each. If the extra stems don't sound cleaner to you, the simpler split is the better default.
How does Ezstems use the Spleeter AI model to separate audio?
Ezstems uses Spleeter, an open-source music source-separation model, as the engine behind a simple upload-and-download workflow. You choose how many stems you want (2, 4, or 5) and an output format, upload a file, and the site processes it server-side with Spleeter before returning the separated tracks. The page notes that processing happens in a queue and that you should not close the browser while it runs, so the separation itself is done on Ezstems' infrastructure rather than in your own browser.
What the stem counts mean in practice
- 2 stems: Typically vocals and accompaniment. Best when you mainly want an instrumental or a cappella track.
- 4 stems: Usually vocals, drums, bass, and "other." A common choice for remixing or practicing a single part.
- 5 stems: Adds a piano category, useful for piano-heavy material where piano otherwise gets buried in the "other" stem.
Output options shown on the page range from MP3 at various bitrates to WAV, with cutoff settings that affect how much high-frequency content is retained. Higher-quality WAV and higher cutoffs preserve more detail but produce larger files.
How the separation works, in plain terms
Spleeter is a deep-learning model trained to recognize and isolate different sound sources in a mixed recording. Ezstems handles the model, the compute, and the file handling; you interact only with the upload form and the download links. The trade-off is control: you get preset stem counts and format options, not fine-grained tuning of the separation itself.
Practical scenario
Suppose you're a vocalist who wants a backing track. Upload the song, pick 2 stems, choose a WAV output if you plan to edit further, and download the accompaniment. If the result has artifacts, try 4 or 5 stems and mute or discard the vocal stem instead—sometimes isolating vocals separately gives a cleaner instrumental.
Limits and when to consider alternatives
The free tier is described as having a 20 uploads-per-day limit and a 20MB file size cap. Paid subscription options are mentioned for bypassing those limits and for API access. If you need batch processing, long files, or programmatic integration, the API access path is the relevant one. For occasional single-song separation, the free tier is usually enough.
A useful next step: test the same short clip at 2 and 4 stems before committing to a full-length file. That tells you quickly whether the extra stem count improves your specific use case. If you want to compare approaches, Ezstems is the tool discussed here; for background on the underlying model, the Spleeter project itself is the primary reference.
DigitalOcean
User reviews (0)