Turn a theme, memory, scene, or emotional idea into original lyrics. The theme is the only required field, and the advanced options let you steer language, title, duration, structure, rhyme, reference direction, and extra notes.
AI Lyrics Generator
Describe the story, scene, or central idea you want the lyrics to express.
Words or images you would like the lyrics to include naturally.
Choose a common style or enter your own.
Choose a common mood or enter your own.
Complete the settings and generate. Your result will appear here without leaving the page.
See how theme, style, structure, and rhyme settings change the shape of a draft before you commit to a direction.
Builds the story from one specific place and an unsaid regret, then returns the chorus to the same memory.
[Verse]
The last ferry cuts across the old light
The night wind files away how we have been
[Chorus]
If the neon still remembers that night
Would it say the words I kept inside
[Bridge]
The tide pulls back and leaves one direction
Shorter lines, stacked images, and a tighter rhyme scheme leave room for the rhythm of a rap section.
[Verse]
The first train cuts the morning, I write my target closer
Not waiting on a lucky streak, I answer every shoulder
[Rap]
Fold the doubt into a pocket, hold the hours in my hand
Every fall recalibrates the centre where I stand
[Chorus]
Before the city wakes, I am already moving
Moves the emotion through scenery, gesture, and omission instead of stacking classical vocabulary.
[Verse]
One lamp beyond the long pavilion, frost on your sleeve
The letter I never sent goes back into the old bag
[Chorus]
If next spring the water still runs to the east
Let one returning boat ask after you for me
These are format samples that show how theme, style, structure, and rhyme settings organize a draft. Actual output changes with your input.
Pick a direction and it loads straight into the workspace. The theme is still yours to write, and every setting stays editable.
Clean storytelling, an easy chorus, and plain language that most listeners can follow.
Pushes rhythm, attitude, word density, and a continuous rhyme scheme.
Organizes imagery through scene, object, and understatement rather than ornament.
Builds a first structure for a Cantopop story and Cantonese phrasing.
Suits everyday detail, first-person narration, and spoken phrasing.
The same lines can come from a chat model, a rhyme dictionary, or a tool that mixes lyrics with audio. What changes is how much control you keep over structure and rhyme.
A full draft with verse, chorus, and bridge labels, plus fixed levels for duration, structure, and rhyme.
Suitable for:Getting a complete, singable structure with controllable rhyme
Continuous text where you supply the section plan and rhyme scheme yourself, and rework it across turns.
Suitable for:Writers who already have a fixed structure and will iterate in conversation
Single words or single-line candidates rather than a whole song.
Suitable for:Replacing one word in a draft you have already finished
Lyrics and a listenable track produced inside one flow.
Suitable for:Hearing a finished result quickly without checking the structure first
Takes finished lyrics on to style, duration, and model settings for an instrumental and vocal.
Suitable for:Turning a signed-off lyric into an arrangement and a vocal
Writing lyrics often stalls before the first useful line. You may know the story and the emotion, yet still struggle to fix a point of view, build a chorus, or keep the imagery consistent from verse to verse. This AI lyrics generator gives that early idea a practical song structure. Describe the theme in ordinary language, such as a reunion on a Hong Kong ferry, a confident post-breakup anthem, or a quiet night drive, and the tool prepares a complete draft you can edit rather than starting from silence.
The workspace keeps the essential decision in front of you and moves optional controls into a collapsed advanced group. The theme is the only fixed requirement. Keywords, music style, mood, output language, title, duration, song structure, rhyme approach, reference direction, and extra instructions are all part of the form, so the whole brief is visible from the start. Style, mood, and language accept your own wording as well as the built-in suggestions.
Every generation uses the same NanoBanana Pro account and the same points balance as the site’s image and video tools. The workspace asks the service for the current point price before it creates a task and shows the authoritative charge afterwards. The balance refreshes when the task is created and again when it finishes, whether it succeeds or fails.
Each item below describes the workspace as it stands now. Optional fields stay open, and the advanced group starts collapsed.
The only field the tool insists on. Describe the story, the scene, or the central idea in ordinary language.
Comma-separated words or pictures the draft should work in naturally.
Pop, Cantopop, Rock, Folk, R&B, Hip-hop / Rap, Electronic, Jazz, or Chinese style, or type a direction of your own.
Warm, Romantic, Nostalgic, Healing, Hopeful, Sad, Passionate, Calm, Playful, or your own description.
English, Simplified Chinese, Traditional Chinese, Cantonese, Japanese, Korean, Spanish, or French.
Leave it empty and the draft arrives with a working title you can replace.
Short (about 1 minute), Medium (about 3 minutes), or Long (about 5 minutes) sets how long the draft should read.
Auto, a four-section layout, the eight-section verse–pre-chorus–chorus–bridge form, verse–chorus–rap–chorus, or a short-form verse–chorus layout.
Natural rhyme, strong rhyme, no emphasis on rhyme, or let the tool decide.
Describe an era, a vocal character, or a writing quality. The field is not a slot for requesting the lyrics of existing songs.
This tool runs on the same NanoBanana Pro account and points balance as the site’s image and video tools. An account is required so the workspace can load the current tool configuration and point price from the service.
See how points workAn AI lyrics generator is a focused writing assistant that turns creative direction into song-ready text. It differs from a general chatbot box because the input is organized around the decisions that shape a lyric: narrative theme, recurring images, genre conventions, emotional intensity, language, section order, and rhyme density. The output is still a draft for human judgement. You decide which lines sound true, which images deserve to repeat, and where the syllable count has to change to fit a melody.
The tool is useful for ideation, co-writing, demo placeholder text, short social content, songwriting exercises, and translation-aware drafting. A producer can test several chorus angles before a session, a newer writer can see how verses and bridges do different jobs, and a content team can put together a short campaign lyric without opening a full music-production workspace.
Originality improves when the brief carries concrete detail instead of a request to imitate a famous song. Use places, objects, actions, contradictions, and emotional turns: the last tram, an unread message, celebrating while still missing home. Before anything is published or recorded, review the result for factual claims, cultural nuance, pronunciation, and fit with the audience you have in mind.
Begin with meaning, add only the controls you need, then treat the result as your own draft.
Write the central story, relationship, scene, or message. Concrete imagery and an emotional turn give the draft a clearer narrative arc.
Choose or type a style, mood, and language. Open the advanced group when duration, structure, rhyme, title, or reference direction matters.
Confirm the point price, create the task, then copy the finished lyrics and rework wording, rhythm, repetition, and singability.
The creative controls appear at once, and the point price is confirmed before anything is generated.
Only the theme is required, so a passing idea can become a draft without a long setup.
Style, mood, and language suggestions speed up entry and still accept wording that is not on the list.
Duration, section structure, and rhyme levels help align the text with a demo or an arrangement plan.
Pick the lyrics language from the presets in the form or type another one, then have a fluent speaker check idiom and pronunciation.
The workspace requests the current price before creating a task and reports the charge that was actually applied.
Verses, choruses, and bridges arrive in a readable panel with a copy button, so they move straight into your editing workflow.
Signed-in results for this tool stay in this browser, up to 12 entries, and can be reviewed or cleared at any time.
Each preset writes a style, mood, structure, and rhyme combination into the form so you can compare directions quickly.
Specific briefs that can actually be sung produce better drafts than broad requests.
Write “a reunion on the last ferry” instead of “a sad love song”, so the draft has something concrete to return to.
Say whether the song speaks in first person, second person, or as an observer, and keep that voice stable across verses.
Give one line that should repeat, or ask the chorus to circle a single question or image.
In the extra notes, state roughly how many syllables a line should carry and where the pauses belong.
An era, a vocal character, or a writing quality is fair to describe. Asking for the lyrics of an existing song or artist is not.
Independent songwriters use the generator to compare narrative perspectives before committing to a melody. Producers prepare temporary lyrics for demos when they need to test arrangement energy quickly. Musicians working across languages use it to explore a draft in another language, then work with a fluent editor to restore natural phrasing. Teachers can turn a topic into a lyric exercise and ask students to critique imagery, rhyme, section function, and emotional progression instead of treating the first output as finished work.
Marketing and social teams can draft short, memorable lines for campaign concepts, event recaps, or creator briefs. Game and film teams can sketch diegetic songs for fictional worlds, while podcast and video creators can explore intro or outro lyric ideas. In each case the reliable workflow is generate, judge, edit, and read aloud. Generated text removes blank-page friction. Taste, ownership, performance, and final quality stay with the writer.
Lyrics to music
This page writes and organizes text. It does not output an instrumental or a vocal take. Once you have finished editing, take the lyrics into the site’s AI music workspace and set the style, duration, and model parameters there.
Writing the lyrics first reduces repeated input and keeps verses, choruses, and bridges clear when you arrange them later. The capabilities and point price that actually apply are the ones shown in the live configuration on the music model page you open.
Treat the result as a working draft. Before you publish or record, check that the lines match your own experience and values, then read or sing them aloud and fix any phrasing, stress, rhyme, or pronunciation that does not sit right. Lyrics written in another language should be reviewed by someone fluent in it and familiar with the cultural context.
Do not ask the tool to reproduce an existing song or an artist’s distinctive writing. Commercial projects also need to be checked against the site terms, the licence of whichever music model you use next, brand requirements, and any other applicable rights. Generated text does not replace legal advice for a specific project.
Start with the story you want to tell, and reach for the advanced controls only when the brief needs them.
Open the lyrics workspace