---
title: ACE-Step v1.5 Base | Runware Docs
url: https://runware.ai/docs/models/ace-step-v1-5-base
description: Open-source music generation with voice cloning, lyric editing, and multilingual support
---
# ACE-Step v1.5 Base

![ACE-Step v1.5 Base](https://assets.runware.ai/covers/ace-step-v1-5-base.jpg)

ACE-Step v1.5 Base is an open-source music generation foundation model built on a hybrid LLM planner and Diffusion Transformer architecture. It generates full tracks from text prompts with support for voice cloning, lyric editing, remixing, cover generation, and compositions up to 10 minutes. It supports over 50 languages and runs on consumer hardware with under 4GB VRAM.

- **Status**: live
- **ID**: `runware:ace-step@v1.5-base`
- Try in Playground
- **Release Date**: January 1, 2026
- **Zero Data Retention**: Yes
- **Capabilities**: Text to Audio, Audio to Audio, Checkpoint

## Pricing

$0.0045 from, per run · ~13s per request

**Measured runs**

- duration 60 · steps 40: `$0.0045` (~4s)
- duration 30 · steps 100: `$0.00675` (~5s)
- duration 60 · steps 100: `$0.009` (~6s)
- duration 60 · steps 300: `$0.0223` (~13s)
- duration 300 · steps 100: `$0.0335` (~25s)

**4 example generations**

- [text-to-audio](https://runware.ai/docs/models/ace-step-v1-5-base/examples#example-ceramic-heat-battery-explainer-song) · 60s: `$0.0039` (~13s)
- [audio-extend](https://runware.ai/docs/models/ace-step-v1-5-base/examples#example-coastal-lighting-retrofit-explainer-narration): `$0.0087` (~22s)
- [audio-repaint](https://runware.ai/docs/models/ace-step-v1-5-base/examples#example-oyster-farm-insurance-explainer-voiceover): `$0.0115` (~30s)
- [audio-driven](https://runware.ai/docs/models/ace-step-v1-5-base/examples#example-avalanche-beacon-training-narration): `$0.0163` (~32s)

billed by compute time, so a request costs what it took · times vary, measured end to end · [How pricing works](https://runware.ai/docs/models-api/pricing)

## Compatibility & Validation

When `inputs.audio` is provided, `duration` cannot be used.

---

When `inputs.audio` is provided, `steps` must be at least `10`.

---

When `negativePrompt` is provided, `CFGScale` must be greater than `1`.

---

When `settings.repaintingStart` or `settings.repaintingEnd` is provided, `inputs.audio` is required.

---

When `strength` is provided, `inputs.audio` is required.

## Request Parameters

**API Options**

Platform-level options for task execution and delivery.

### [taskType](https://runware.ai/docs/models/ace-step-v1-5-base#request-tasktype)

- **Type**: `string`
- **Required**: true
- **Value**: `audioInference`

Identifier for the type of task being performed.

### [taskUUID](https://runware.ai/docs/models/ace-step-v1-5-base#request-taskuuid)

- **Type**: `string`
- **Required**: true
- **Format**: `UUID v4`

UUID v4 identifier for tracking tasks and matching async responses. Must be unique per task.

### [outputType](https://runware.ai/docs/models/ace-step-v1-5-base#request-outputtype)

- **Type**: `string`
- **Default**: `URL`

Audio output type.

**Allowed values**: `URL` `base64Data` `dataURI`

### [outputFormat](https://runware.ai/docs/models/ace-step-v1-5-base#request-outputformat)

- **Type**: `string`
- **Default**: `MP3`

Specifies the file format of the generated output. The available values depend on the task type and the specific model's capabilities.

- \`MP3\`: Compressed audio, smaller file size.
- \`WAV\`: Uncompressed, high-quality audio.
- \`FLAC\`: Lossless compression.
- \`OGG\`: Open-source compressed audio format (Vorbis codec).

**Allowed values**: `MP3` `WAV` `FLAC` `OGG`

### [audioSettings](https://runware.ai/docs/models/ace-step-v1-5-base#request-audiosettings)

- **Path**: `audioSettings.bitrate`
- **Type**: `object (3 properties)`

Audio encoding settings for controlling the bitrate, number of channels, and sample rate of the generated audio. Only applicable for lossy output formats (`MP3` and `OGG`). When using lossless formats (`WAV` or `FLAC`), this parameter must not be provided.

The available sample rates and valid bitrate ranges depend on the output format. For `OGG`, bitrate limits also vary by the number of channels.

**MP3 bitrate limits**:

Bitrate limits for MP3 are the same regardless of mono or stereo.

| Sample Rate | Min Bitrate | Max Bitrate |
| --- | --- | --- |
| 8,000 Hz | 8 kbps | 64 kbps |
| 11,025 Hz | 8 kbps | 64 kbps |
| 12,000 Hz | 8 kbps | 64 kbps |
| 16,000 Hz | 8 kbps | 160 kbps |
| 22,050 Hz | 8 kbps | 160 kbps |
| 24,000 Hz | 8 kbps | 160 kbps |
| 32,000 Hz | 32 kbps | 320 kbps |
| 44,100 Hz | 32 kbps | 320 kbps |
| 48,000 Hz | 32 kbps | 320 kbps |

**OGG bitrate limits — Mono (1 channel)**:

| Sample Rate | Min Bitrate | Max Bitrate |
| --- | --- | --- |
| 8,000 Hz | 8 kbps | 40 kbps |
| 12,000 Hz | 16 kbps | 48 kbps |
| 16,000 Hz | 16 kbps | 96 kbps |
| 24,000 Hz | 16 kbps | 80 kbps |
| 48,000 Hz | 32 kbps | 224 kbps |

**OGG bitrate limits — Stereo (2 channels)**:

| Sample Rate | Min Bitrate | Max Bitrate |
| --- | --- | --- |
| 8,000 Hz | 16 kbps | 80 kbps |
| 12,000 Hz | 16 kbps | 96 kbps |
| 16,000 Hz | 24 kbps | 192 kbps |
| 24,000 Hz | 32 kbps | 160 kbps |
| 48,000 Hz | 48 kbps | 256 kbps |

> [!NOTE]
> **Lossless formats**: When `outputFormat` is set to `WAV` or `FLAC`, the `audioSettings` parameter is not available since these formats produce uncompressed or lossless audio with no configurable encoding settings.

#### [bitrate](https://runware.ai/docs/models/ace-step-v1-5-base#request-audiosettings-bitrate)

- **Path**: `audioSettings.bitrate`
- **Type**: `integer`
- **Min**: `8`

Audio bitrate in kbps.

#### [channels](https://runware.ai/docs/models/ace-step-v1-5-base#request-audiosettings-channels)

- **Path**: `audioSettings.channels`
- **Type**: `integer`
- **Default**: `2`

Number of audio channels. 1 for mono, 2 for stereo.

**Allowed values**: `1` `2`

#### [sampleRate](https://runware.ai/docs/models/ace-step-v1-5-base#request-audiosettings-samplerate)

- **Path**: `audioSettings.sampleRate`
- **Type**: `integer`

Audio sample rate in Hz.

### [webhookURL](https://runware.ai/docs/models/ace-step-v1-5-base#request-webhookurl)

- **Type**: `string`
- **Format**: `uri`

Webhook URL that receives JSON responses via HTTP POST when generation tasks complete. For batch requests with multiple results, each completed item triggers a separate webhook call as it becomes available.

**Learn more** (1 resource):

- [Webhooks](https://runware.ai/docs/models-api/webhooks) (platform)

### [deliveryMethod](https://runware.ai/docs/models/ace-step-v1-5-base#request-deliverymethod)

- **Type**: `string`
- **Default**: `sync`

Determines how the API delivers task results.

**Allowed values**:

- `sync` Returns results directly in the API response. Over HTTP, a task still running after 120 seconds continues as an async task and comes back with status 'processing'. Poll for its results using getResponse.
- `async` Returns an immediate acknowledgment with the task UUID. Poll for results using getResponse.

**Learn more** (1 resource):

- [Task Polling](https://runware.ai/docs/models-api/task-polling) (platform)

### [uploadEndpoint](https://runware.ai/docs/models/ace-step-v1-5-base#request-uploadendpoint)

- **Type**: `string`
- **Format**: `uri`

Specifies a URL where the generated content will be automatically uploaded using the HTTP PUT method. The raw binary data of the media file is sent directly as the request body. For secure uploads to cloud storage, use presigned URLs that include temporary authentication credentials.

**Common use cases:**

- **Cloud storage**: Upload directly to S3 buckets, Google Cloud Storage, or Azure Blob Storage using presigned URLs.
- **CDN integration**: Upload to content delivery networks for immediate distribution.

```text
// S3 presigned URL for secure upload
https://your-bucket.s3.amazonaws.com/generated/content.mp4?X-Amz-Signature=abc123&X-Amz-Expires=3600

// Google Cloud Storage presigned URL
https://storage.googleapis.com/your-bucket/content.jpg?X-Goog-Signature=xyz789

// Custom storage endpoint
https://storage.example.com/uploads/generated-image.jpg
```

The content data will be sent as the request body to the specified URL when generation is complete.

### [ttl](https://runware.ai/docs/models/ace-step-v1-5-base#request-ttl)

- **Type**: `integer`
- **Min**: `60`

Time-to-live (TTL) in seconds for generated content. Only applies when `outputType` is `URL`.

### [includeCost](https://runware.ai/docs/models/ace-step-v1-5-base#request-includecost)

- **Type**: `boolean`

Include task cost in the response.

### [numberResults](https://runware.ai/docs/models/ace-step-v1-5-base#request-numberresults)

- **Type**: `integer`
- **Min**: `1`
- **Max**: `4`
- **Default**: `1`

Number of results to generate. Each result uses a different seed, producing variations of the same parameters.

**Inputs**

Input resources for the task (images, audio, etc). These must be nested inside the \`inputs\` object.

### [audio](https://runware.ai/docs/models/ace-step-v1-5-base#request-inputs-audio)

- **Path**: `inputs.audio`
- **Type**: `string`

Audio input (UUID or URL).

**Core Parameters**

Primary parameters that define the task output.

### [model](https://runware.ai/docs/models/ace-step-v1-5-base#request-model)

- **Type**: `string`
- **Required**: true
- **Value**: `runware:ace-step@v1.5-base`

Send `auto` instead of an AIR identifier to have the model router choose the model for this task. The rest of the request is unchanged, every value you send is kept, and only the parameters you leave out are filled in. See [Model Router](https://runware.ai/docs/models-api/model-router).

**Learn more** (3 resources):

- [Text To Image: Model Selection](https://runware.ai/docs/learn/text-to-image#model-selection) (learn)
- [Image Inpainting: Model Specialized Inpainting Models](https://runware.ai/docs/learn/image-inpainting#model-specialized-inpainting-models) (learn)
- [Image Outpainting: Other Critical Parameters](https://runware.ai/docs/learn/image-outpainting#other-critical-parameters) (learn)

### [positivePrompt](https://runware.ai/docs/models/ace-step-v1-5-base#request-positiveprompt)

- **Type**: `string`
- **Required**: true
- **Min**: `2`
- **Max**: `3000`

Text prompt describing elements to include in the generated output.

**Learn more** (1 resource):

- [Prompts](https://runware.ai/docs/learn/prompts) (learn)

### [negativePrompt](https://runware.ai/docs/models/ace-step-v1-5-base#request-negativeprompt)

- **Type**: `string`
- **Min**: `2`
- **Max**: `3000`

Prompt to guide what to exclude from generation. Ignored when guidance is disabled (CFGScale ≤ 1).

**Learn more** (1 resource):

- [Prompts: Negative Prompts](https://runware.ai/docs/learn/prompts#negative-prompts) (learn)

### [duration](https://runware.ai/docs/models/ace-step-v1-5-base#request-duration)

- **Type**: `float`
- **Min**: `30`
- **Max**: `300`
- **Step**: `0.1`
- **Default**: `60`

Length of the generated audio track in seconds.

### [seed](https://runware.ai/docs/models/ace-step-v1-5-base#request-seed)

- **Type**: `integer`
- **Min**: `0`
- **Max**: `2147483647`

Random seed for reproducible generation. When not provided, a random seed is generated in the unsigned 32-bit range.

### [steps](https://runware.ai/docs/models/ace-step-v1-5-base#request-steps)

- **Type**: `integer`
- **Min**: `1`
- **Max**: `300`
- **Default**: `100`

Total number of denoising steps. Higher values generally produce more detailed results but take longer.

**Learn more** (1 resource):

- [Steps](https://runware.ai/docs/learn/steps) (learn)

### [CFGScale](https://runware.ai/docs/models/ace-step-v1-5-base#request-cfgscale)

- **Type**: `float`
- **Min**: `1`
- **Max**: `30`
- **Step**: `0.01`
- **Default**: `10`

Guidance scale representing how closely the output will resemble the prompt. Higher values produce results more aligned with the prompt.

**Learn more** (1 resource):

- [Cfg Scale](https://runware.ai/docs/learn/cfg-scale) (learn)

### [strength](https://runware.ai/docs/models/ace-step-v1-5-base#request-strength)

- **Type**: `float`
- **Min**: `0`
- **Max**: `1`
- **Step**: `0.01`
- **Default**: `0.5`

Fraction of steps using the input source instead of generated output.

**Learn more** (3 resources):

- [Image To Image: Strength The Transformation Intensity](https://runware.ai/docs/learn/image-to-image#strength-the-transformation-intensity) (learn)
- [Image Inpainting: Strength Controlling Transformation Intensity](https://runware.ai/docs/learn/image-inpainting#strength-controlling-transformation-intensity) (learn)
- [Image Outpainting: Strength Understanding The Critical Threshold](https://runware.ai/docs/learn/image-outpainting#strength-understanding-the-critical-threshold) (learn)

**Settings**

Technical parameters to fine-tune the inference process. These must be nested inside the \`settings\` object.

### [bpm](https://runware.ai/docs/models/ace-step-v1-5-base#request-settings-bpm)

- **Path**: `settings.bpm`
- **Type**: `integer`
- **Min**: `30`
- **Max**: `300`

Beats per minute. If not set, the model decides automatically.

### [guidanceType](https://runware.ai/docs/models/ace-step-v1-5-base#request-settings-guidancetype)

- **Path**: `settings.guidanceType`
- **Type**: `string`
- **Default**: `apg`

Controls how guidance is applied during generation.

**Allowed values**:

- `apg` Adaptive Projected Guidance.
- `cfg` Classifier-Free Guidance.

### [keyScale](https://runware.ai/docs/models/ace-step-v1-5-base#request-settings-keyscale)

- **Path**: `settings.keyScale`
- **Type**: `string`

Musical key and scale in '{Note}{Accidental} {Mode}' format (e.g. 'C major', 'F# minor', 'Bb major').

### [lyrics](https://runware.ai/docs/models/ace-step-v1-5-base#request-settings-lyrics)

- **Path**: `settings.lyrics`
- **Type**: `string`
- **Min**: `10`
- **Max**: `3000`

Song lyrics, typically formatted like a lyrics website.

### [repaintingEnd](https://runware.ai/docs/models/ace-step-v1-5-base#request-settings-repaintingend)

- **Path**: `settings.repaintingEnd`
- **Type**: `float`
- **Min**: `0`
- **Max**: `300`

End time in seconds for repaint region. Requires input audio. Values beyond audio duration append new audio.

### [repaintingStart](https://runware.ai/docs/models/ace-step-v1-5-base#request-settings-repaintingstart)

- **Path**: `settings.repaintingStart`
- **Type**: `float`
- **Min**: `-300`
- **Max**: `300`

Start time in seconds for repaint region. Requires input audio. Negative values prepend audio before the start.

### [timeSignature](https://runware.ai/docs/models/ace-step-v1-5-base#request-settings-timesignature)

- **Path**: `settings.timeSignature`
- **Type**: `integer`

Beats per measure. If not set, the model decides automatically.

**Allowed values**: `2` `3` `4` `6`

### [vocalLanguage](https://runware.ai/docs/models/ace-step-v1-5-base#request-settings-vocallanguage)

- **Path**: `settings.vocalLanguage`
- **Type**: `string`
- **Default**: `en`

ISO 639-1 language code for vocals. `unknown` for instrumental or auto detection.

**Allowed values**: `unknown` `ar` `az` `bg` `bn` `ca` `cs` `da` `de` `el` `en` `es` `fa` `fi` `fr` `he` `hi` `hr` `ht` `hu` `id` `is` `it` `ja` `ko` `la` `lt` `ms` `ne` `nl` `no` `pa` `pl` `pt` `ro` `ru` `sa` `sk` `sr` `sv` `sw` `ta` `te` `th` `tl` `tr` `uk` `ur` `vi` `yue` `zh`

## Response Parameters

### [taskType](https://runware.ai/docs/models/ace-step-v1-5-base#response-tasktype)

- **Type**: `string`
- **Required**: true
- **Value**: `audioInference`

Identifier for the type of task this response belongs to.

### [taskUUID](https://runware.ai/docs/models/ace-step-v1-5-base#response-taskuuid)

- **Type**: `string`
- **Required**: true
- **Format**: `UUID v4`

UUID v4 identifier echoed from the original request, used to match async responses to their tasks.

### [audioUUID](https://runware.ai/docs/models/ace-step-v1-5-base#response-audiouuid)

- **Type**: `string`
- **Required**: true
- **Format**: `UUID v4`

UUID of the output audio.

### [audioURL](https://runware.ai/docs/models/ace-step-v1-5-base#response-audiourl)

- **Type**: `string`
- **Format**: `uri`

URL of the output audio.

### [audioBase64Data](https://runware.ai/docs/models/ace-step-v1-5-base#response-audiobase64data)

- **Type**: `string`

Base64-encoded audio data.

### [audioDataURI](https://runware.ai/docs/models/ace-step-v1-5-base#response-audiodatauri)

- **Type**: `string`
- **Format**: `uri`

Data URI of the output audio.

### [seed](https://runware.ai/docs/models/ace-step-v1-5-base#response-seed)

- **Type**: `integer`

The seed used for generation. If none was provided, shows the randomly generated seed.

### [cost](https://runware.ai/docs/models/ace-step-v1-5-base#response-cost)

- **Type**: `float`

Task cost in USD. Present when `includeCost` is set to `true` in the request.

## Examples

### text-to-audio (text-to-audio)

[Listen to audio](https://assets.runware.ai/examples/ace-step-v1-5-base/0df33ff8-b87a-428c-9bcc-8308883aea3c.mp3)

```json
{
  "taskType": "audioInference",
  "taskUUID": "7fc96bd8-41bc-4da9-a007-c7edfbdb7696",
  "model": "runware:ace-step@v1.5-base",
  "positivePrompt": "Create a complete original English explainer song for a B2B film about thermal batteries in ceramic manufacturing. Blend warm documentary narration with restrained melodic singing: an assured female alto delivers the technical lines clearly, then opens into a memorable sung refrain. Organic industrial chamber-folk production with prepared piano, muted marimba, bowed upright bass, brushed frame drum, soft ceramic clicks, and a low kiln-like air hum. Begin sparse and tactile, build gently as the process becomes clear, and finish with a calm, resolved mnemonic hook. Intelligent, practical, quietly optimistic, and credible rather than glossy. Keep every lyric intelligible, with natural phrasing and enough space for the explanation. No hard sell, no exaggerated climate claims, no EDM, no synthwave, no funk, no brass fanfare, and no crowd vocals.",
  "duration": 60,
  "settings": {
    "lyrics": "[Spoken Intro]\nA kiln remembers every flame.\nNow it can remember the wind and sun, too.\n\n[Verse]\nWhen renewable power is plentiful,\nelectric heaters warm a core of refractory brick.\nLayer by layer, heat is held inside,\nquietly waiting behind insulated steel.\n\n[Spoken]\nWhen production calls, controlled air moves through the core,\ncarrying stored heat back to the kiln.\nNo combustion in the battery.\nNo change to the clay.\nJust dependable heat, ready for the next firing.\n\n[Refrain]\nBank the heat when clean power flows.\nBring it back when the kiln door closes.\nHold the hours. Shape the flame.\nMake tomorrow from the same old clay.\n\n[Outro]\nFor tiles, tableware, and technical ceramics:\nsteady temperatures, lower fuel demand,\nand energy that works around the production day.\n\n[Sung Tag]\nBank the heat.\nShape what comes next.",
    "guidanceType": "apg",
    "bpm": 112,
    "timeSignature": 4,
    "keyScale": "C minor",
    "vocalLanguage": "en"
  }
}
```

---

### audio-driven (audio-driven)

[Listen to audio](https://assets.runware.ai/examples/ace-step-v1-5-base/fcd078d0-32a6-4129-9c36-29f535f90d77.mp3)

```json
{
  "taskType": "audioInference",
  "taskUUID": "d419ae88-b5e9-4c2e-ac78-f93f7de89aa7",
  "model": "runware:ace-step@v1.5-base",
  "positivePrompt": "Transform the supplied musical source into a newly arranged Nordic electroacoustic documentary bed for a professional avalanche-rescue e-learning lesson. Preserve the source's recognizable five-note motif, harmonic movement, and measured pulse, but reorchestrate them with felt piano, bowed vibraphone, low cello harmonics, subtle granular wind textures, and a muted frame drum. The mood is focused, reassuring, and cold-clear rather than dramatic. Use spacious, modern public-broadcast production with restrained dynamics and seamless transitions. Keep the music beneath the voice and gently duck it during key instructions. Add one calm, authoritative adult female instructor speaking natural English with precise diction, moderate pacing, and no sung delivery. Narration, spoken verbatim: \"Before leaving the trailhead, switch every transceiver to send and complete a group check. If an avalanche occurs, move to a safe position, call emergency services, and switch every rescuer's beacon to search. Follow the direction and distance display quickly at first. Inside ten meters, slow down. Below three meters, keep the beacon close to the snow and use a careful cross pattern to find the lowest reading. Probe perpendicular to the surface. When you strike the victim, leave the probe in place and begin shoveling from downhill.\" End with a brief, resolved instrumental reprise of the source motif.",
  "negativePrompt": "Singing, chanting, choir, theatrical acting, alarmist tone, action-movie percussion, loud music over speech, distorted narration, slurred words, excessive reverb, abrupt ending.",
  "strength": 0.65,
  "settings": {
    "guidanceType": "apg"
  },
  "inputs": {
    "audio": "https://assets.runware.ai/assets/inputs/1f61c8cd-e262-4a82-b4eb-6c4b0d6f1f76.mp3"
  }
}
```

---

### audio-extend (audio-extend)

[Listen to audio](https://assets.runware.ai/examples/ace-step-v1-5-base/f9bc8197-ef98-4dd4-b5f5-d6b7af37a7c1.mp3)

```json
{
  "taskType": "audioInference",
  "taskUUID": "e4a8aa82-b8fb-418e-b75d-e7ea6246ee97",
  "model": "runware:ace-step@v1.5-base",
  "positivePrompt": "Prepend a refined instrumental introduction that develops naturally into the supplied coastal infrastructure explainer. Match the source exactly in key, tempo, meter, instrumentation, mix perspective, and production character: A minor, 78 BPM, restrained industrial-documentary scoring with muted marimba, warm analog bass, brushed percussion, soft bowed cello harmonics, distant harbor air, and occasional understated buoy-bell color. Begin sparsely with harbor ambience and a single low marimba motif, then gradually introduce bass pulse, cello texture, and brushed rhythmic detail. Keep the mood calm, credible, modern, and quietly optimistic rather than cinematic or dramatic. Shape the final measures so the harmony, groove, ambience, loudness, and room tone flow seamlessly into the existing opening downbeat. The new introduction must remain instrumental, leaving the first spoken line of the source untouched and clearly intelligible when it begins.",
  "strength": 0.7,
  "settings": {
    "repaintingStart": -12,
    "repaintingEnd": 0,
    "guidanceType": "apg"
  },
  "inputs": {
    "audio": "https://assets.runware.ai/assets/inputs/34c9c90e-b245-425a-bd40-91f696560181.mp3"
  }
}
```

---

### audio-repaint (audio-repaint)

[Listen to audio](https://assets.runware.ai/examples/ace-step-v1-5-base/c1d36e65-1c43-4717-83bc-d69b2ca9f9a2.mp3)

```json
{
  "taskType": "audioInference",
  "taskUUID": "96513e32-4d82-40b0-8fa8-d6d0bc025e61",
  "model": "runware:ace-step@v1.5-base",
  "positivePrompt": "Replace only the designated interior region of the existing oyster-farm insurance explainer. Preserve the source narrator's calm female voice, intimate documentary delivery, tempo, loudness, room tone, and restrained prepared-piano score. Speak this copy clearly and verbatim, never sung: “Traditional coverage waits for an adjuster. Tidelock watches the water instead. When verified salinity or storm-surge thresholds threaten your lease, payment starts automatically—no damage inspection, no paperwork delay—so crews can reset cages, replace seed, and get the next tide working.” Create a noticeable but brand-compatible change beneath the narration: gradually submerge the existing piano into a tactile underwater soundscape of filtered buoy knocks, tiny shell clicks, hydrophone bubbles, and one warm bass swell. Keep every word intelligible and natural. Enter seamlessly from the preceding music, let the underwater texture bloom in the middle, then surface gradually and reconnect exactly with the following prepared piano, percussion, ambience, and narration. No abrupt edits, new intro, ending cadence, jingle, chorus, or announcer-style exaggeration.",
  "strength": 0.75,
  "settings": {
    "repaintingStart": 8,
    "repaintingEnd": 18,
    "guidanceType": "cfg"
  },
  "inputs": {
    "audio": "https://assets.runware.ai/assets/inputs/d394af1a-68ad-40ac-a4d2-ee625d2585ce.mp3"
  }
}
```