Make All Your Songs the Same Volume (Free, In Your Browser, No Uploads)

Yes — you can make every song in a folder play at the same volume, in your browser, with nothing uploaded. A browser mastering pass levels each file to −14 LUFS integrated with a −1.0 dBTP true-peak ceiling and re-measures the result after render, so the numbers are verified, not promised. One song at a time is free forever, no account. A whole folder takes the bulk queue — unlocked once, from $1.99, kept forever.

Why is every song in your collection a different volume?

Because they were mastered in different decades, by different people, to different standards — or by no one at all. A loudness-war-era commercial master can sit several dB louder than a CD rip from the late 80s. Live recordings, Bandcamp downloads, and voice memos land wherever they land. And AI-generated tracks are their own scatter: we measured 12 AI-generated exports from our test library and got a median of −15.2 LUFS integrated with a range from −16.4 all the way up to −12.3 — a spread of over 4 dB inside a single batch of exports. Put any of these next to each other in a playlist and you become the human normalizer, riding the volume knob between every track.

Why did the old fixes — MP3Gain, Sound Check, Normalize — stop working?

Search this problem and the top answers are genuinely old. MP3Gain was a clever tool — it losslessly rewrites the gain field inside each MP3 frame — but its last stable release shipped in 2005, it only handles MP3s, it can only move volume in fixed 1.5 dB steps, and it predates the LUFS standard streaming platforms now use. iTunes Sound Check adjusts playback level inside Apple's own players (toward roughly −16 LUFS) but never touches your files — burn them, share them, or play them anywhere else and the mismatch is right back. And Audacity's classic Normalize effect aligns peaks, not loudness — Audacity's own docs draw exactly this distinction. Two files with identical peaks can still differ by several dB in perceived volume, which is why a folder of "normalized" files still lurches.

Verified July 2026 — tool statuses per the linked sources.

What is the difference between peak and loudness, in 60 seconds?

Peak is the single tallest sample in the file — an instant. Loudness (LUFS) is a gated average of the energy your ear actually integrates over the whole track. Your ears track the average, not the instant, which is why peak-normalizing a quiet acoustic song and a dense rock track to the same ceiling leaves them sounding wildly different. It is also the number streaming platforms level toward: Spotify normalizes playback to −14 LUFS. The platform-by-platform numbers are in our streaming loudness targets guide, and the peak side of the story is in the true peak explainer.

How do you level one song, free?

Drop the file into the TrackGleam masterer. It analyzes first — integrated LUFS, true peak, loudness range, crest factor, stereo width, plus BPM and musical key with its Camelot code — so you can see exactly how far off the file is before anything changes. Then it renders a master targeting −14 LUFS integrated with a −1.0 dBTP ceiling (BS.1770-4 gated) and measures the finished render, so the report shows what is actually in your file. A/B it against the original at matched streaming volume, then download the WAV or MP3. No watermark, no account, and the audio never leaves your device — open your browser's devtools network tab while it runs and watch nothing upload.

Just want to diagnose before you fix? The free LUFS target checker and loudness penalty checker will measure any file and tell you how far from streaming level it sits.

How do you level 30 songs at once?

This is the part the old tools never solved and the new tools put behind subscriptions. TrackGleam's bulk queue takes as many files as you drop on it — 30, 100, whatever the folder holds — and lets you preview one track free. Unlocking the whole queue works like this, plainly: any one-time credit purchase, starting at $1.99, unlocks bulk forever. Once unlocked, "Master all" runs the free engine on every track in the batch with no per-track cost — a 20-track batch levelled to a consistent, verified −14 LUFS costs $1.99 total, once, ever. If you want the AI-Enhanced treatment instead — per-track AI decisions, linear-phase finishing EQ, per-band stereo width — "Master all with AI" spends one credit per track ($1.99 for 1, $5 for 10, $20 for 100, $99 lifetime with a 50-per-month allowance). Everything is one-time: no subscription, no account, and there is a 14-day money-back guarantee.

We will not pretend the batch button is free — it is not. But it is the closest thing to a modern MP3Gain that exists: pay once, level folders forever, in any format the browser decodes, to an actual measured standard instead of 1.5 dB guesses.

Which playlists is this actually for?

Wedding or party mix. The one place a volume lurch is unforgivable — nobody is standing at the laptop between the dinner set and the first dance. Level the whole folder once and the night runs itself.

DJ set prep. Consistent gain staging before you ever touch the trim knobs — and the analysis card reads BPM and Camelot key on every track, which is half your set planning done while it levels.

A folder of AI generations. Suno and Udio exports drift several dB between generations. Batch-levelling is step one; if the tracks also sound thin or brittle, start with the Suno mastering guide, and muddy Udio exports have their own fix guide.

Spoken word and voice. Interviews, sermons, narration — level plus cleanup matters more than mastering polish there, so use the voice cleanup path with neural noise removal.

Ambient, lofi, and sleep playlists. Quiet genres should not be slammed to pop loudness — the genre presets include ambient- and lofi-appropriate quieter targets, so "consistent" does not have to mean "loud."

Why does "nothing uploads" matter for a batch?

One song, one upload — fine. Thirty WAVs is gigabytes. Server-side tools have to ingest all of it before the first render starts, and your unreleased music sits on someone else's machine while they do. Local processing starts decoding the instant you drop the folder, works the same on hotel wifi as on fiber, and the files never exist anywhere but your device. That is a checkable claim, not a slogan: the network tab is right there.

What can't volume-levelling fix?

Being honest about the limits: levelling matches loudness, not character. If one track was mixed muddy and the next mixed bright, they will be equally loud and still sound like different rooms — that is a tonal problem, which full mastering can narrow but a great mix fixes properly. Clipping that is already printed into a file cannot be un-clipped; the limiter can stop it getting worse, not rewind it. And if you are cutting vinyl or shipping a label release, a human mastering engineer making per-track judgment calls is still the right spend — no $1.99 batch button replaces that, ours included.

Level a track free — no signup, nothing uploads