Do you have to sing your own songs to call yourself a singer-songwriter?

Sing the line before you argue about the label

Hold the last word of this for four beats:

And I stayed where the light couldn't reach

Now the same line, one word swapped:

And I stayed where the light couldn't go

Nine syllables either way. The meaning barely moves. But the first version parks you on reach: a narrow, tense vowel, and then a consonant your tongue has to assemble before you can let the note die. Most people clip it at beat two and hope the snare covers them. The second version drops you onto an open oh that you can hold until your air runs out, with nothing to close.

A writer who sings their own material makes that swap without deciding to. It happens the second time through, when the throat complains and the hand moves. A writer pitching the song to an artist may never make it, because the note is not theirs to hold and nothing in the room tells them it hurts.

That reflex, not the phrase on the press release, is what this question is actually about. So the answer first, and the rest of this is the defence of it: no, you do not need to sing well, and you do not strictly need to sing at all. What you need is a voice in the writing. If no throat anywhere constrained a single line, the label is a sticker.

Three jobs that keep getting mixed up

They look similar from outside because all three end with a song. They differ in one thing: where the constraints come from, and who eats the loss when the song does not work.

The songwriter writing for other people. Constraints come from a stranger's range, a stranger's vocabulary, and usually a stranger's manager. You write wide, because you do not know where their break sits. You write transferable, because the singer has to be able to own the detail. When the song fails, it fails in a pitch meeting and you never learn why. The trade is real and it pays differently; if that is the direction, songwriting as a paid job is a separate set of rules.

The singer who performs other people's songs. Constraints come from the arrangement and from the audience's memory of the original. This is interpretation, which is a craft with its own difficulty, and it is not writing. Nobody in this role is confused about the label, which is why the argument never happens here.

The singer-songwriter. Constraints come from one body. Your range decides where the chorus sits. Your lung capacity decides how long the lines get. Your embarrassment decides which words never make it in. When the song fails, it fails in a room with forty people in it, and you find out immediately.

Notice that none of these is defined by talent. They are defined by which limits reach the page.

One piece of honesty before this goes further, because some readers arrive at articles like this wanting a finished recording at the end. This site writes words. The lyrics generator costs 20 credits, which is twenty cents, and it hands back a title and a lyric. It does not sing. There is no audio, no vocal, no mp3 anywhere in it. To hear the thing you take the words to a service that makes sound, and Suno is the usual choice. Suno is a separate company and we have nothing to do with it, do not resell it, and do not control its rules.

What changes in the text when you write for your own throat

This is the part people skip, because it sounds like production and it is actually writing. Five things move.

Range decides where the payoff word goes. The chorus has one word that carries it. If that word lands at the top of your range, you have written a song you can only perform on good days. Writers who sing learn to keep the chorus within about a sixth of the verse, then get the lift from density and arrangement instead of from altitude. The song sounds bigger without going higher. That is not a compromise, it is a technique, and it is why so many quiet writers sound enormous on record.

Breath decides line length. Count syllables to the next place you can inhale, not to the next comma. Sixteen syllables with no gap reads fine on paper and strands you in the booth. The fix is almost never cutting the idea. It is moving one word across the bar line so there is a hole to breathe in.

Vowels decide what you can sustain. Open vowels stay open by themselves. Narrow ones need holding, and the consonant after them is usually what kills the note. Any note you intend to hold for more than a beat should land on a word you can still be singing in two seconds without your jaw locking. This is why so many choruses across every genre resolve on go, away, stay, slow. It is not a poverty of vocabulary. It is anatomy leaking into the lyric sheet.

Consonant clusters decide tempo. The streets stretched past is four words and a pile-up. At 80 bpm it is fine. At 140 it turns to mush and you will start dropping consonants to survive, which means the audience stops hearing the line you were proud of. Writers who sing test the fast lines at tempo before they finish them. Writers who do not, do not.

Shame decides vocabulary, and nobody admits this one. There are words you cannot say with a straight face into a microphone. For some people it is baby. For others it is forever, or soul, or their own city's name. It is not that the word is bad. It is that your face does something when you sing it, and the audience reads your face before it parses the line. A singer-songwriter deletes those words early and quietly. A writer supplying an artist leaves them in, because the artist might carry them fine.

Here is the thing about all five: they are audible. A listener who knows nothing about vocal technique can hear a song written by somebody who has sung it fifty times, because the lines sit in the mouth. They cannot tell you why. They just do not stumble.

Why a song built for somebody else's voice is built differently

Go the other way and the constraints invert.

You write wider, on purpose. Not knowing the singer's break means leaving the melody room to be transposed without wrecking the shape. A melody that only works in one key is a melody that only one person can buy.

You write detail that transfers. I still have your father's coat is specific and anyone can sing it. I still have Ray's coat is more specific and dead on arrival, because the artist now has to sell a person they have never met. This is the constraint a singer-songwriter simply does not have. Your own life is pre-authorised. That is the single biggest advantage of the role and most people who hold it waste it by writing generically anyway, which is a shame, because the analysis of any song you love usually turns up one detail nobody could have invented.

You are careful with the delivery you cannot control. A line that lands only when it is half-whispered will get belted by somebody, somewhere, and then it is just a sentence. So the emotion has to be in the words rather than parked in the performance. Which, and this matters for the end of this article, is exactly the discipline that generated vocals now force on everyone.

You watch the demo trap. Demo vocals get imitated more than anybody admits. Phrasing you invented because of your own limits, a note you hedged, a syllable you rushed because you were out of air, all of it can end up copied into the finished record by a singer who has no such limit. If you demo your own songs for other people, sing them straighter than feels natural.

And the blunt test that settles most arguments: put the lyric in a stranger's mouth. Read it out loud in someone else's accent, in the wrong gender, at the wrong age. If it collapses, you wrote a singer-songwriter song. That is not a fault. It just means the pitch meeting is not where it belongs.

You write strong and sing weak. Five exits, and one of them is best.

This is the real question hiding under the identity question, so it gets a real answer instead of a menu.

1. Sing it anyway, and write to your range on purpose. This is the right answer for most people and I will defend it. A narrow range is not a defect, it is a constraint, and constraints have been the making of more records than technique has. Practical version: find your comfortable sixth, write the whole chorus inside it, put the peak word on an open vowel, and thin the arrangement out under the peak so nothing is competing with you. The song gets its size from dynamics and from the words, not from the top note. If you are early enough that this all sounds abstract, starting from the first song is the better entry point than fixing a voice you have not used yet.

2. Stop singing some of it. Spoken verses, half-sung, rap, near-monotone over a moving chord. Entire genres run on this. It is not a workaround for people who cannot sing, it is a delivery choice with its own demands, mainly that the words have to be much better because nothing melodic is covering them.

3. Bring in a singer and settle the split before the session, in writing. Not after, not verbally, not on the strength of everyone being friends. The number of collaborations that die over an unspoken assumption about credit is the most boring tragedy in music.

4. Generate the vocal. Cheap, fast, and it changes what you are. See the next section, which is the one I would not skip.

5. Stop performing entirely and write for other people. Different career, different math, different skill. The writing gets harder in the ways described above and easier in others.

Where exit 1 fails, honestly: some songs genuinely need a belt at the top and will not survive being written down to fit you. Transpose first. If it still needs the belt, the song is asking for someone else, and holding onto it out of pride is how good songs die in notebooks.

What is left of the role when a machine does the singing

Here is where the title stops being a quibble and starts costing something.

A generated vocal takes away everything that lived in the performance. Timbre is not yours. Phrasing is not yours. The breath that catches before the last line is not yours, the crack on the high note is not yours, the decision to push the third syllable instead of the second is not yours. Every one of those is a place where singers put meaning that is not in the text, and you no longer have any of them.

What remains is the words, the shape, the point of view, the specific detail, and the decision about what the song refuses to say. That is a smaller role. It is also, and I mean this without softening it, the harder half. Delivery can rescue a mediocre lyric for about three minutes. Nothing rescues a lyric that has nothing in it when the voice singing it is competent, pleasant, and completely uninvested.

So the honest position: if a model sings it, you are the author of the words. Call yourself a songwriter, which is a title that has been honourable for a hundred years and does not need inflating. Calling yourself a singer-songwriter when no human sang anything is the kind of claim that gets checked exactly once, in public, and never recovers.

Three things change in the writing itself the moment the vocal is generated, and they are worth doing before anyone asks you to:

• Cut every line that only works if somebody cries it. You will recognise them. They are the ones that look thin on the page and you have been telling yourself the delivery will carry.
• Move the emotional information into nouns. Adverbs and adjectives were always a request for the singer to do the work. The kitchen light was still on at four beats I was desperately lonely in any voice, and it is the only one of the two that survives a neutral reading.
• Resolve the hook in words, not on a held note. If the payoff of your chorus is the length of the last vowel, a generated vocal will hand you back something technically correct and emotionally flat. The chorus has to close on meaning, not on duration.

There is a small illustration of all this sitting in our own tool, and it is the reason this article exists. The lyrics generator has a vocal field with three options, and no duet among them. It also produces no sound whatsoever. Which means that setting can only do one single thing: change the words. Pick a different vocal and you do not get a different singer, you get different line lengths, different vowels, different density. That is the whole argument of this article, compressed into a dropdown. The voice was always in the text. It is just that until recently you could hide it behind an actual throat.

The rest of the tool works the same way, by the way: one free field of 300 characters for what the song is about, then closed lists for structure, style out of 53 curated ones, one to three emotions out of twelve, point of view, audience age, language. Twenty cents a run, and the current prices are on the pricing page. Nothing in it makes audio.

When this whole framing is wrong

Two cases, because a position that never fails is not a position.

If you are writing for theatre or for a specific commissioned performer, none of the above applies cleanly. You are writing for a known voice that is not yours, which is a fourth job with its own rules, and the constraints come from someone else's body measured in advance.

And if you are writing only to put a song in front of one person, a wedding, a funeral, a birthday, the label question is noise. Nobody in that room cares what you call yourself. They care whether the second verse says something true about a person they know. Spend the argument-energy there instead.

So: do you have to sing them?

No. You have to write as though a body has to get through it.

That is the whole distinction, and it holds in every direction. Sing your own songs badly and write to your limits, and the label fits. Hand them to a better singer and keep writing for a throat, it still fits, because the constraint is still doing the work. Generate the vocal and write like the words are the only thing left standing, because they are, and then call yourself a songwriter and stop worrying about a hyphen.

The people who lose this argument are the ones writing lines nobody could sing and no machine could save, then arguing about what to put on the poster. If you want a place to start that is not this argument, pick a song you love and take it apart line by line, or read the two or three books that working writers keep recommending to each other. Then write something with a breath mark in it.