HiAPI
OverviewModel MarketplaceAPI KeysUsage StatisticsCall LogsBillingReferralPlaygroundStorageChangelogContact UsSettings
Display unit
N
Powered by hiapi
Settings

Welcome

Contact Us

Lyria 3.5 turns prompts or lyrics into complete songs for songwriting, video soundtracks, and music ideas. Try it online and view pricing and API examples.

Provider: Google

Category: music generation

Endpoint: /v1/tasks

Status: Available

Cost: See live page pricing

Back to Models

Lyria 3.5

New
by GoogleMusic

Lyria 3.5 turns prompts or lyrics into complete songs for songwriting, video soundtracks, and music ideas. Try it online and view pricing and API examples.

Pricing

Standard Usage

180 Credits/ song

Input

0 / 10000

More parameters

Generation settings

0

Use section tags such as [Verse] and [Chorus]. Provide lyrics or a music description; At least 1 characters

Set a title for this song

Beats per minute, at least 1. Leave empty for an unspecified tempo

Requested duration, 1–240 seconds. Actual audio length may vary; leave empty to omit; Range: 1 - 240, Integer only

Advanced

Supplied as text
Estimated cost180 Credits / song

Output

Ready for generation

Configure your parameters and click "Run" to see the output here.

--

Lyria 3.5 generation example

Start with a morning music description to generate an acoustic folk track led by fingerpicked guitar and soft brushed drums.

Instrumental folk · Acoustic guitar

Morning acoustic

Fingerpicked guitar with soft brushed drums creates a warm, quiet morning atmosphere.

Prompt

Warm instrumental indie folk, fingerpicked acoustic guitar, soft brushed drums, quiet morning.

Related audio models

Choose an adjacent route for music, single-speaker speech, or multi-speaker dialogue.

Structured songs

MiniMax Music 1.5

Generate complete Chinese or English songs from a style brief and lyrics.

Auto lyrics and instrumentals

MiniMax Music 2.6

Choose custom lyrics, automatic lyrics, or instrumental generation.

Multi-speaker dialogue

ElevenLabs Text to Dialogue

Assign a voice to each line for podcasts, drama, and character dialogue.

Compare AI audio APIs

Model specifications

Inputs
Music description or lyricsOptional title, tempo, and duration
Output
One audio track per request
Duration
Requested: 1-240 seconds
Billing
180 Credits / song

How to use Lyria 3.5

Use three steps to organize the input, validate the audio, and integrate the API.

01 Describe the music

Provide a music brief, lyrics, or both; at least one is required.

02 Add optional controls

Optionally add a title, BPM, and requested duration to guide the song.

03 Integrate the API

Reuse the model ID and input object to create an async task and retrieve the audio result.

API Quickstart

Copy a minimal request. See the documentation for complete parameters, task retrieval, and errors.

Endpoint
POST /v1/tasks
Model ID
lyria-3.5
cURL
curl -X POST "https://api.hiapi.ai/v1/tasks" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "lyria-3.5",
  "input": {
    "bpm": "90",
    "length": 30,
    "prompt": "Warm instrumental indie folk, fingerpicked acoustic guitar, soft brushed drums, quiet morning."
  }
}'

The create request immediately returns a taskId; audio generation continues asynchronously.

Read the full API documentation

LYRIA 3.5

Lyria 3.5: AI music and song generation

Google DeepMind’s music generation model brings melody, lyrics, vocals, and arrangement together in structured, expressive compositions.

Text / lyric creationInstruments and vocalsGenre and tempo
Explore the model and its work ↓
Lyria 3.5 model cover artwork

About Lyria 3.5

Lyria 3.5 is Google DeepMind’s AI music generation model for songs and instrumental compositions created from text. Genre, mood, instrumentation, and vocal delivery establish a creative direction that the model develops into melody, accompaniment, and song sections.

Lyria 3.5 advances melodic richness, understanding of lyric structure, and vocal expression. Its music can include contrasts between verses and choruses, changes in instrumental layers, and a more nuanced emotional performance alongside the central musical theme.

Its range spans acoustic folk and pop through R&B, electronic music, and cross-genre arrangements. Multilingual singing and creative control over tempo and duration support songwriting, video soundtracks, programme themes, and musical exploration.

From musical ideas to complete expression

Lyria 3.5 combines melody, lyrics, and vocal performance with varied genres, instrumentation, and dynamics to create complete songs.

01

Coherent melody and full-song structure

Lyria 3.5 combines melody, harmony, rhythm, and accompaniment into songs with verses, choruses, and bridges. Melodic ideas can develop across sections as the arrangement grows from a sparse opening into a fuller sound, giving the composition room for progression and contrast.

Richer melodic structures are a focus of this model generation. A pop chorus opening up, the instrumental layers of a folk arrangement, or the changing intensity of an electronic track offer different ways to develop a musical theme and carry an emotional thread through a song.

02

Lyrics shaped around musical structure

Lyria 3.5 supports songs built around a theme as well as compositions based on existing lyrics. Characters, scenes, and emotions in the text can provide a narrative thread, with verses developing the story and choruses concentrating its central theme.

Creative applications

Ads and short video

Build a music direction with a clear mood, rhythmic entrance, and ending for branded clips.

Podcasts and shows

Prepare themes for intros, outros, and transitions while leaving room for narration.

Games and interactive content

Explore looping, narrative, or character-theme arrangements around a scene and mood.

Independent drafts

Quickly explore melody, lyrics, vocals, and instrumentation before detailed production.

Start shaping your musical direction

Use the Playground to explore Lyria 3.5’s musical expression with a music description or lyrics.

Open Playground
Need to compare models and billing tiers?View live pricing for all models

Frequently asked questions

Can it generate a vocal song without lyrics?

Yes. Describe the theme, vocal language, and genre so the model can shape the music and vocal performance, including the song lyrics.

How should existing lyrics work with a genre request?

Put the existing words in lyrics, and put genre, instruments, and mood in prompt. At least one of lyrics or prompt is required; when both are present, they guide lyrics and musical direction separately.

How can I create an instrumental backing track?

State instrumental or no vocals in the text description and specify the instruments, rhythm, and atmosphere you want. For example: an instrumental acoustic-guitar backing track with soft brushed drums, warm and gentle.

How should I express verses, choruses, and bridges?

Use section markers such as [Verse], [Chorus], and [Bridge] in lyrics, and describe the emotion and movement of each section. For example, let the verse develop the story, repeat the central line in the chorus, and change the mood in the bridge to create contrast.

Why are BPM and duration separate?

BPM describes speed and groove, while duration describes how long the piece should run. Requested duration guides the overall development and is not a hard cut on the generated result.

How can I compare different arrangement versions?

Keep the lyrics and theme the same, then change one element at a time, such as replacing acoustic guitar with piano or changing drum intensity. When using a seed, keep the other inputs consistent for comparison; a fixed seed still does not guarantee identical results.

Improved lyric instruction following and structural awareness place more emphasis on the relationship between sections. A recurring chorus, a narrative verse, and a bridge that changes the emotional direction can work with the melody and vocal performance to turn written material into a song.

03

More nuanced vocals and emotional delivery

The vocal improvements in Lyria 3.5 focus on naturalness, pronunciation, and emotional expression. Singing can take on gentle, bright, low, or powerful qualities, with tone and delivery creating different degrees of intimacy and intensity.

The relationship between a voice and its accompaniment also shapes a song’s character. Sparse acoustic instrumentation brings the singer and lyrics forward, rhythm-led R&B gives phrasing a different role, and layered harmonies add space and depth to a chorus. These elements can be combined within a single musical composition.

04

Multilingual songs across musical styles

Lyria 3.5 generates songs across multiple languages and musical styles, including pop, folk, electronic music, and soul. The vocal language, instrumental arrangement, and rhythm work together to shape a composition’s sound, extending beyond a simple change of lyric text.

French pop, Hindi acoustic songwriting, and Latin dance music illustrate this range of musical directions. The cadence of a language interacts with the beat, while light accompaniment, gentle delivery, or a stronger rhythmic emphasis suits different themes and emotional settings.

05

Cross-genre arrangements and instrumentation

Lyria 3.5 can bring different genres and instruments into one composition. Acoustic guitar, keyboards, bass, brass, and electronic textures can take melodic, harmonic, or rhythmic roles, producing arrangements with light, warm, retro, or distinctly electronic qualities.

Combinations such as bossa nova with trip-hop, or jazz harmony with Caribbean rhythms, illustrate the possibilities of cross-genre music. A shared beat, harmonic foundation, and sectional structure give the different elements a relationship and keep the composition centred on a musical idea.

06

Tempo, dynamics, and musical duration

Lyria 3.5 offers creative control over tempo and duration, helping music take different rhythmic and expressive forms. A slower pace creates space, a lively groove brings movement, and duration influences how sections unfold, when a theme appears, and how a piece concludes.

Dynamics and arrangement density add further variation. Sparse instrumentation can support a voice or narration, building drums and harmonies can increase emotional intensity, and light electronic textures can sustain a detailed rhythmic flow. These qualities lend themselves to songs, programme music, and video soundtracks.