How Long Should a Music Video Be? Platform Lengths, Song Length, and What to Make First

Is there a right length for a music video?

The short answer: There is no single correct length. Match the platform and the job: a full music video for YouTube runs the length of the song, usually three to four minutes, while short-form clips for TikTok, Reels and Shorts perform best at 15 to 60 seconds regardless of how much the platform technically allows. The more useful question is what to make first. A 30-second cut is almost always the right first artifact, even for a three-minute song, because it tells you whether the visual direction is working before you commit to the whole track. Length is not a free dial — it maps directly onto shot count, render time and cost. Melodious prices video generation at 20 credits per second, so a three-minute video costs six times a 30-second one, and roughly 92% of that spend is the motion rather than the stills.

"How long should it be" sounds like a question about taste. It isn't. It's a question about three things that have real answers: where the video is going, whether the video's job is to be the song or to sell it, and what you're willing to spend before you know the look works.

Most advice answers only the first. That's the least interesting one, because the platform limits are generous and mostly irrelevant — nobody's music video failed because it was under the cap. The decisions that matter are the other two, and they pull in opposite directions: the version that serves the song wants to be long, and the version that finds an audience wants to be short. If you're new to the whole workflow, how to make an AI music video covers the full pipeline; this post is about the one number that dictates most of the others.

What each platform actually wants

Start with the distinction that does the most work: the cap is not the target. Every short-form platform allows far more than you should use, and the number that matters is the length at which clips actually perform, which is much shorter.

One caveat on those caps: they move. YouTube Shorts was 60 seconds until October 2024 and is three minutes now; TikTok and Instagram have both raised their limits more than once and allow well past the Shorts figure. Any number printed in a blog post — including this one — is a snapshot, so check the platform's own help pages before you plan against a cap. Design for the performing length instead and you'll never be near one.

The performing length is short for a structural reason: short-form feeds reward completion and replays. A 20-second clip that loops cleanly on your hook can be watched three times in the space one 90-second clip gets abandoned. That is arithmetic on watch time rather than a matter of taste — and it's why making a music video for TikTok is a genuinely different craft problem rather than a crop of the long version. The same logic applies to shape: see the aspect ratio guide for the dimensions each of these wants.

Match the song, or cut it down?

This is the real decision, and it's binary more often than people expect.

Match the song when the video is the artifact — the YouTube upload, the thing you link from your release announcement, the version that sits on your site. Here a video that stops at 90 seconds of a three-minute track feels broken. The listener arrived to hear the song; ending early breaks the contract.

Cut it down for everything else, which is most things. A promotional clip has one job: make someone want to hear the rest. That job is done in the chorus, or the one line people would sing back, and everything before it is you spending attention you haven't earned yet.

The mistake is treating these as versions of one video with a slider between them. They're different edits with different structures:

If you're only ever going to make one, make the cut-down. An artist with a strong 25-second clip and no full video is in a better position than one with a full video nobody has seen — you can always extend later, and the short one is what does the finding.

There is a third length, and it's the one most artists never make: the 60-to-90-second sizzle. Not a full video, not a feed clip — a condensed version that shows the whole visual idea in the time someone will actually give it. That's the version you put in an electronic press kit, send to a playlist curator or a booking agent, and run as paid promotion, because all three of those audiences are evaluating you rather than enjoying you. It's structured like a trailer: a moment from each section of the song, ending on the strongest one. If you're releasing a single and can only fund two edits, the sizzle and the 15-second hook cover more ground than a full video and nothing else.

Why 30 seconds is the right first version

Here's the part that has nothing to do with platforms: even when you know you want a full three-minute video, the first thing you should produce is about 30 seconds of it.

The reason is that you can't evaluate a direction from a description. You write "late-night city, neon reflections, a lone singer" and you have a mental image; the model has a different one. Somewhere between those two is the actual video, and you find out where by looking at real frames. A 30-second cut is enough to answer the questions that matter — is the palette right, does the character look like who I meant, is the pacing too fast, does this feel like my song? — and it answers them at a fraction of what the full version costs.

This is why Melodious's automatic first run always produces a 30-second clip regardless of how long your song is. It's a deliberate choice, not a limitation. Pressing send commits you to a spend, so the first run commits to the smallest amount worth looking at, and it's the one length available on every plan. A three-minute auto-run would spend six times as much before you'd seen a single frame.

Think of it the way a director thinks about a test shoot. Nobody shoots the whole video to find out whether the lighting concept works. You shoot one setup, look at it, and then commit. The 30-second version is that test — except the test is the deliverable, because a 30-second cut is exactly what TikTok, Reels and Shorts want anyway. You're not throwing it away.

What length actually costs

Length is not a free parameter, and this is where the honest version of the advice diverges from the generic version. Every extra second of video is another rendered clip. There is no shared work between shots, so cost scales linearly — no bulk discount, because there's no bulk.

Melodious prices video generation at 20 credits per second (100 credits = $1.00):

The ratio in those last two columns is the most useful thing on this page. Seeing stills is cheap. Rendering motion is not. For a 30-second clip, the keyframe images cost around 48 credits — roughly 8% of the total — and the video render accounts for the other ~92%.

Three consequences follow from that ratio. Seeing a full set of keyframes before you render costs about an eighth of the video, so there is no excuse for rendering a direction you haven't looked at. Regenerating one keyframe that came out wrong costs 7 to 13 credits depending on how long that shot is; regenerating a rendered clip costs twelve times that. And the expensive decision isn't "how long" — it's "am I ready to render at all". Which is why in Melodious the video render is the one step that still requires an explicit click: the keyframes run automatically from your first message, and nothing turns into video until you approve it.

One thing that has changed and is worth being precise about: the storyboard is no longer something you get for free. Audio analysis, chatting, and editing a plan you already have are not charged. But on a project's first message the storyboard and its keyframe images are produced in the same run, so getting a storyboard at all now means paying the image-stage cost — about 48 credits, stated under the composer before you send. Older pages that say planning is free predate that change. For context on what any of this replaces, what a music video costs puts the traditional numbers next to these.

Length is really a shot-count decision

Duration and shot count aren't independent. Pick a length and you've picked, roughly, how many shots you're committing to — because pacing determines how long each shot runs, and the length divided by that is your shot count.

On the automatic first run you don't pick from that table — Melodious reads your song's tempo and picks for you: roughly 118 BPM and up gets the fast packing, roughly 90 BPM and down gets the slow one, and everything between gets the mix. Ask in the composer if you want a different one. Scale it up and the arithmetic holds: a two-minute video is four times the shots at the same pacing. Pacing changes how many scenes you get, not what the video costs — every pacing packs to the same total length, and credits are priced on duration, not shot count.

The practical consequence: a longer video is a longer shot list, and a longer shot list is more places for the look to drift. Fifteen shots give consistency fifteen chances to fail — a different face, a different palette, a location that doesn't match. That's the real argument against jumping straight to full length, more than the money. Get one recognizable world working across four shots before you ask for fifteen. Storyboarding against the song's sections is where you make that call, and the consistency choices that make it cohere matter more as the shot count climbs.

Common mistakes when choosing a length

Every one of those mistakes comes from treating length as the first decision. It's the last one: it falls out of where the video is going and how confident you are in the look, and both of those are things you can settle before you pick a number.

How this works in Melodious

The length decision has moved to after you've seen something, rather than before.

Attach a song, type what you want to see, and press send — that one message produces a storyboard and generates all the keyframe images for it. There's no interview, no length picker, and nothing to approve before you see images. The credit cost is stated under the composer before you send, because pressing send is what commits the spend. Then:

1. The first run is 30 seconds. Not derived from your song — fixed, whatever the track is. It's the only length available on every plan and the cheapest thing worth judging, so it's what the automatic run always builds. (What your song does decide is the cut speed: tempo sets whether those 30 seconds arrive as roughly four long shots or seven short ones.)
2. Look at the keyframes. This is the evaluation the whole approach is built around — real frames in your style, for about 8% of what the finished clip would cost.
3. Change the length by asking. Say "make it 60 seconds" in the composer and the storyboard is rebuilt at the new duration. It's an edit, not an answer to a question you were asked upfront.
4. Approve the render explicitly. Turning keyframes into video is the ~92% step and it has its own click. Melodious will never start a render because you sent a chat message.

The storyboard stays yours throughout — edit any scene's prompt, add or remove scenes, regenerate a single keyframe that missed. Nothing is locked because it was written automatically. Longer durations do need a Pro plan or the Single Release Video pack; the 30-second clip is available on every plan, which is part of why it's the default. The video duration docs have the tier-by-tier detail, and pricing has the current numbers.

The one-line version

Make 30 seconds first, whatever your song's length. It's the cheapest artifact that answers the only question you can't answer from a description — does this look right? — and it's simultaneously the exact length that TikTok, Reels and Shorts want. Then extend to the full song for the YouTube version, once you're extending something you've already seen work rather than gambling on one.

If you're still deciding what shape the thing should be as well as how long, vertical or horizontal first is the companion decision — and it's worth making both before you generate anything, not after.

Start with a 30-second version of your song — one message gets you a storyboard and keyframes, and you decide about length once you can see what you're deciding about.

Frequently asked questions

How long should a music video be?+

It depends on where it's going. A full music video for YouTube or your own site runs the length of the song — typically three to four minutes. Short-form clips for TikTok, Instagram Reels and YouTube Shorts perform best at 15 to 60 seconds, built around one hook, even though all three platforms allow far longer. If you're making one thing, make the short one first: it costs less, and it tells you whether the look works before you commit to the full track.

Should a music video be the same length as the song?+

Only for the definitive version — the one that lives on YouTube and on your own site, where people arrive expecting to hear the whole track. Everywhere else, a music video is promotional and should be cut down. Matching the song is the exception, not the default: most of the places your video actually gets discovered reward a fraction of it.

How long should a music video be for TikTok or Instagram Reels?+

15 to 60 seconds, and closer to 15–30 for a music clip. Every one of those platforms allows far longer than that — YouTube Shorts caps at three minutes, and TikTok and Instagram have raised their own limits repeatedly — but the cap is not the target, and the caps move often enough that you should check the platform's own help page rather than any blog post. Short-form feeds reward completion and replays, and a 20-second clip that loops cleanly on the hook outperforms a 90-second one that gets swiped past.

Why does Melodious make a 30-second video first?+

Because it's bounded spend on an unproven direction. A 30-second clip costs 600 credits; a three-minute one costs 3,600. Until you've seen a storyboard and keyframes in your actual style, you don't know whether the direction is right — so the first run commits to the smallest amount worth looking at. You ask for a longer version once the short one has told you something.

How much more does a longer music video cost?+

It scales linearly. Melodious charges 20 credits per second of finished video, so 30 seconds is 600 credits, 60 seconds is 1,200, two minutes is 2,400, and a three-minute video is 3,600. There's no discount for length, because there's no shared work between shots — every extra second is another rendered clip.

Can I change the length after the video has been storyboarded?+

Yes. The storyboard isn't locked because it was written automatically — ask for a different length in the composer and the plan is rebuilt at the new duration. That's the whole reason for starting short: judging a length against a storyboard you can see is a much easier call than picking one before anything exists.

Start with 30 seconds, then decide

Attach your song, say what you want to see, and you get a storyboard and keyframes for a 30-second cut. Go longer once you've seen whether the direction is right.

Make a 30-second version