Music GenerationActive
Text to Music v2 Lyrics (ElevenLabs)
Write a section-by-section composition plan with lyrics and style cues. ElevenLabs Music v2/v2.5 renders it into a finished MP3 song.
Overview
Text to Music v2 Lyrics is an ElevenLabs music generation model for creating full songs from structured lyrics. You provide a JSON composition plan with timed sections, then the model sings the lyrics and builds a matching arrangement. It outputs a single mastered audio track, which is useful when you need predictable song structure and timing.
What you can build
- Short vocal hooks for ads, trailers, and social clips
- Verse/chorus song drafts from your own lyrics
- Multilingual vocal songs, when your lyrics are written in another language
- Section-based arrangements with clear transitions, like intro → verse → chorus → bridge
- Instrumental breaks by leaving a section’s lyric text empty
Inputs
- A composition plan you paste as JSON text. It contains an ordered list of sections (up to 30) that generate one after another.
- For each section, you provide: - A section label and lyric text in one block, with lines separated by newlines. The label typically appears in brackets at the top.
- A target length for that section in milliseconds.
- A list of style phrases to include, such as genre, instrumentation, vocal type, tempo, or key. These style phrases should be written in English.
- An optional list of style phrases to avoid.
- An optional setting that controls how strongly the section should stay consistent with surrounding sections.
- A requested MP3 quality setting. You choose the sample rate (22.05 kHz or 44.1 kHz) and bitrate (32–128 kbps).
Outputs
- One MP3 audio track.
- The track contains the full song arranged in the same order as your planned sections.
- Vocals are generated where you include lyrics. Instrumental-only sections are possible when you omit lyrics for a section.
Recommended settings
- Put your most important genre and mood cues in the first section’s style phrases. Later sections tend to follow that overall direction.
- Use the highest context consistency setting when you want one cohesive song sound across all sections.
- Lower the context consistency setting when you want a clear style shift between sections, like a genre switch at the bridge.
- Keep lyric lines short and easy to sing. Very long lines can sound rushed within fixed section lengths.
Limitations
- The composition plan supports up to 30 sections total.
- Each section must be between 3 seconds and 120 seconds.
- Total song duration must be between 3 seconds and 10 minutes.
- Each section’s lyric block has line limits. Plans with very long lines or inconsistent formatting can fail or produce uneven delivery.
- Low-quality or messy input text can reduce results. Common issues include missing section labels, inconsistent line breaks, or mixing styles and lyrics in the same line.
- Exact reproducibility isn’t guaranteed. The same plan can produce different audio over time as the system updates.
Safety & compliance
- Only submit lyrics you have the rights to use. Don’t paste copyrighted lyrics you don’t control.
- Don’t request content that violates applicable laws or platform rules.
- Commercial usage rights can vary by plan and terms. Review ElevenLabs music terms before releasing or distributing generated tracks.
Example prompts
Great starting points for text-to-music-v2-lyrics.
{"chunks":[{"text":"[Verse]\nMorning light is on the window\nCoffee steam and quiet streets\nEvery road is calling my name","duration_ms":20000,"positive_styles":["upbeat indie pop","jangly electric guitars","warm bass","bright female vocals","120 bpm","great production quality"],"negative_styles":["autotune","heavy distortion"]},{"text":"[Chorus]\nWe are running into summer\nNothing here can slow us down\nHold my hand and feel the thunder\nWe are running into summer","duration_ms":20000,"positive_styles":["big anthemic chorus","layered harmonies","driving drums"]}]}Music Generation
{"chunks":[{"text":"[Intro]","duration_ms":10000,"positive_styles":["cinematic orchestral","soft piano","warm strings","slow 70 bpm","emotional","great production quality"],"negative_styles":["drums","electronic synths"]},{"text":"[Verse]\nThe lanterns fade along the harbor\nI keep your letters in my coat\nThe tide will carry what I promised","duration_ms":25000,"positive_styles":["intimate male baritone vocals","sparse piano","gentle cello"],"context_adherence":"high"},{"text":"[Outro]","duration_ms":10000,"positive_styles":["swelling strings","piano fade out"],"negative_styles":["vocals"]}]}Music Generation
{"chunks":[{"text":"[Verse]\nNeon lights on a rainy avenue\nEngines hum like a heartbeat in the dark\nI drive until the city sleeps","duration_ms":20000,"positive_styles":["80s synthwave","punchy bass arpeggio","gated reverb drums","airy male vocals","110 bpm","great production quality"],"negative_styles":["acoustic guitar","lo-fi"]},{"text":"[Drop]","duration_ms":10000,"positive_styles":["bright synth lead solo","pulsing bass","energetic"],"negative_styles":["vocals"],"context_adherence":"medium"},{"text":"[Chorus]\nMidnight runner, don't look back\nChasing stars along the track","duration_ms":15000,"positive_styles":["catchy synth pop chorus","stacked vocals","wide stereo synths"]}]}Music Generation
{"chunks":[{"text":"[Verse]\nDust is rising on the county road\nOld guitar and a heavy load\nI am heading where the rivers go","duration_ms":20000,"positive_styles":["acoustic folk","fingerpicked acoustic guitar","harmonica","warm female vocals","stomp and clap percussion","95 bpm","great production quality"],"negative_styles":["electronic synths","autotune"]},{"text":"[Chorus]\nCarry me home on the evening light\nSing me a song for the long, long night","duration_ms":15000,"positive_styles":["uplifting folk chorus","group harmonies","banjo"]}]}Music Generation
API quick start
Run text-to-music-v2-lyrics with a single API call.
{
"prompt": "{\"chunks\":[{\"text\":\"[Verse]\\nMorning ligh…",
"outputFormat": "mp3_44100_128"
}curl -X POST "https://api.wiro.ai/v1/Run/elevenlabs/text-to-music-v2-lyrics" \
-H "Content-Type: application/json" \
-H "x-api-key: YOUR_WIRO_API_KEY" \
--data-binary @- <<'JSON'
{
"prompt": "{\"chunks\":[{\"text\":\"[Verse]\\nMorning ligh…",
"outputFormat": "mp3_44100_128"
}
JSONOther models you might like
Explore similar music generation models.
