V3 Is Here: Edit Any Song in One Pass

I just shipped the biggest update ChangeLyric has ever gotten. V3 lets you edit every lyric in a song in a single submission. Change the hook, rewrite verse two, tweak the bridge, rework the outro: all at once, one pass, one download.

If you used the V2 Horizon engine, you know the pain. You had to work section by section, manually setting start and end times for each chunk of audio. It worked, but it was tedious. V3 eliminates that entire workflow.

This is still a pro tool. You still need DAW experience and patience. But the friction between "I want to change these lyrics" and "here are my stems" has been cut dramatically.

What V2 Required (And Why It Was Painful)

The V2 Horizon engine was a major upgrade at the time. It introduced better vocal tone matching and eliminated content filters. But the workflow still required you to process the song in sections.

Want to change lyrics in verse 1 and the chorus? That was two separate submissions. You had to manually set the time range for each section, write new lyrics for just that portion, wait for processing, then do it again for the next section. Then you would stitch the results together in your DAW.

For a full song with changes across four sections, that could mean four separate runs. Each one taking 15-30 minutes to process. That is an hour or more just waiting around, not counting the time spent setting time ranges and managing separate outputs. It got the job done, but I always knew there was a faster way.

How V3 Actually Works

V3 is a completely different architecture. Here is the workflow from start to finish.

Step 1: Upload Your Track

Drop in an MP3 of the song you want to modify. V3 runs it through a song recognition engine to get word-level timing data. This is not basic transcription: it maps every word to its exact position in the audio, which is how the engine knows where to place your new lyrics.

Step 2: Edit the Lyrics

The transcription loads into an editor where you can change any line you want. One line, five lines, every line in the song. No word count matching required. Put whatever you want in each line. The AI orchestrator figures out timing and placement.

This is the biggest quality-of-life improvement. In V2, you were constrained to editing a specific time range. In V3, you see the entire song's lyrics and just change what you want. Edit lines scattered across every section and they all get processed together.

Step 3: Submit and Process

Hit submit and the V3 engine handles the rest. First-run processing takes about 5 minutes for 10 line changes as the engine analyzes your track. Subsequent versions on the same song process faster, roughly 2 minutes per 10 lines changed. Processing continues in the background so you can close the page and come back when it is done.

Step 4: Download Your Stems

You get a ZIP file with everything separated: original vocal stem, instrumental stem, and new vocal tracks for each method you selected. Import into Pro Tools, Logic, Ableton, FL Studio, whatever you use. Full control over levels, EQ, effects, and placement.

Real talk on the vocals inside that ZIP. On tonally distinctive singers, those new vocal tracks sometimes still need to be upscaled by passing them through the voice changer to fully nail the timbre. We are actively building that upscaling step into the V3 experience so it stops being a separate workflow.

Still the Only Platform With Zero Content Filters

Nothing has changed on this front. ChangeLyric remains the only lyric swap platform that does not run your uploads through content moderation or copyright detection algorithms. No audio fingerprinting. No automated rejections. No false positives on your own original productions.

Platforms like Suno and Udio continue to aggressively block copyrighted uploads. I get it: they have legal teams to satisfy. But for professional producers who understand copyright law and handle licensing responsibly, those filters are just friction. They block legitimate work constantly.

I built ChangeLyric for producers who do not need algorithmic babysitting. You decide what you have rights to modify. That philosophy has not changed with V3, and it never will.

Free Trial Reset for V3

If you signed up for ChangeLyric before and used your free trial on V2, good news: V3 is a completely different engine, and I have reset trial eligibility. Even if you already had an account, you can try V3 for free and see the difference yourself.

V3 is such a substantial improvement over the section-by-section workflow that I want everyone to experience it firsthand. The difference in speed and convenience is immediately obvious the first time you use it.

Pro Tips for Getting the Best Results

V3 handles the heavy lifting, but these techniques will help you get cleaner results faster. I have been using all of these on done-for-you client projects since the V3 beta, and they make a real difference.

Tip 1: The Scratch Track Method

After V3 gives you your stems, build a final scratch track in your DAW. Layer the original vocal with whatever sections you replaced using V3. Some parts of the original might sound better left alone, so keep them. The point is to create one continuous vocal comp.

Then export that comp as a single vocal file and run it through the Voice Changer tool. This smooths out the tonal differences between original and AI-generated sections. Instead of having two obviously different vocal characters, you get one unified voice across the whole track.

This technique alone eliminates the most common complaint I hear: "the AI sections sound different from the original." Of course they do. But when you run the entire vocal through the same voice conversion, everything matches. I wrote more about why AI vocals mismatch the original in this deep dive on tonal mismatch.

Tip 2: Isolate the Vocal First

Sometimes you get better results by running the Vocal Splitter before V3. Use it to separate the lead vocal from backing vocals, then upload just the lead vocal stem plus the instrumental to V3.

Why? The V3 engine does its own stem separation internally. But if you feed it a cleaner input (isolated lead vocal with a clean instrumental) it has less work to do and often produces tighter results. This is especially useful for songs with dense backing harmonies that can confuse the AI.

This extra step takes about two minutes and can noticeably improve output quality on complex mixes.

Tip 3: Run Your Final Mix Through Suno for a Full Vocal Replacement

If you want to completely eliminate tonal differences between original and AI-generated vocal sections, this is the most robust method I have found. Take your finished mix (original vocal comped with V3 sections, laid over the instrumental) and upload the full master to Suno as a cover.

Uploading the full master rather than isolated vocals is important because the instrumental helps Suno maintain proper timing throughout the song.

Previously, you could create a persona from about 30 seconds of your target vocal that shows dynamic range, quiet to loud. However, Suno has since disabled personas for audio uploads, removing this vocal cloning avenue. The good news is that ChangeLyric V3 outputs can achieve a similar effect, and we also added built-in custom voice training to the Voice Changer (on all plans) so you can clone a very specific singer that isn't in the 200+ model library.

If you still want to use Suno's cover feature without personas, set the style influence low and the weirdness setting low. Generate at least four variations, then extract the vocals from each using Ultimate Vocal Remover or LALAL.AI and comp the best phrases back into your original instrumental.

This is a more robust alternative to the Voice Changer approach from Tip 1. The Voice Changer works well for smoothing minor inconsistencies. Think of it as a light polish. But Suno re-sings the entire track in one consistent voice, which means every tonal mismatch between original and AI sections disappears completely.

If you are dealing with bigger tonal gaps or want a completely different vocal character, this is the move. I wrote a full breakdown of this workflow with comparison tests if you want the details.

Tip 4: Vote on What Gets Built Next

I built a feature voting board where you can submit and upvote ideas. I actually read every submission, and the most requested features get built first. If something about V3 frustrates you or if there is a workflow improvement that would save you time, that is the place to put it.

Some of the best improvements to ChangeLyric have come directly from user feedback. The Vocal Splitter and Voice Changer tools both started as feature requests. V3 itself was heavily shaped by producers telling me the section-by-section workflow was too slow.

Lyric-Swap Tool vs V2 Horizon: When to Use Which

The lyric-swap tool is the primary product, and it has two modes behind the scenes: Overdub (the new bulk mode, default) and Classic (the per-line mode that used to be called V3). Both modes work from the same editor: upload a song, edit lyrics, submit once, get stems back. Overdub is the right call when you are changing many lines across a whole song in one pass, and Classic is the right call when you only want surgical control over a few lines.

V2 Horizon is a separate, older tool with a different UI and a different use case. You select a specific time range, hand-pick 60-90 seconds of primer audio, and the engine extends just that section. Some producers still prefer it for single-section fixes where matching the original vocal tone on that one chunk matters most. Credits are shared across all of these on your plan, so you can pick whichever fits the project.

Pricing: Same $9/Month

V3 is included in the existing Starter plan at $9/month. No price increase. Credits are shared between V2 Horizon and V3, so you can use whichever engine fits the project.

At $9/month for unmoderated access to full-song lyric swapping, this is a fraction of what our done-for-you service charges for manual work. We bill $200+ per song for the same workflow that V3 now automates in a single submission. If you are a working producer handling multiple projects, this pays for itself on the first song.

Ready to try it? Head to the dashboard and run your first V3 submission. If you are a new user, the free trial gives you enough credits to test it on a real project.

What This Means for the Workflow

The practical impact of V3 is straightforward: less waiting, less busywork, more time actually mixing. A full song edit that used to take four separate V2 submissions and an hour of processing time can now be done in one pass. You spend your time in the DAW where it matters, not babysitting a queue.

This is one piece of what I use internally on 600+ done-for-you client projects, faster and more accessible than the old section-by-section flow. On every service order I still chain V3 with the voice changer, V2 Horizon for polyphonic sections, and Suno or Udio for smoothing where moderation allows. V3 is the bulk-swap base, not the whole pipeline.

It is still not a one-click solution. You still need ears, DAW skills, and willingness to iterate. If you have never worked with vocal stems before, start with the getting started guide to understand the fundamentals. But for experienced producers, V3 removes the biggest bottleneck that existed in the old workflow.

Bottom Line

V3 turns ChangeLyric from a section-by-section tool into a full-song editing engine. Edit every lyric in one submission. Get stems back. Mix in your DAW. No content filters blocking your work.

Whether you are creating custom versions for sync licensing, building demo alternatives for songwriting sessions, or making parodies for content, the output is separated stems you control in your DAW. The days of manually setting time ranges and stitching four separate outputs together are over.

This is a pro tool for pro producers. If that is you, V3 is ready.