Audio Cleaner (Podcast) API - Complete Integration Guide

Overview

Key Features

• Multi-format and multi-channel support: supports text, web pages, YouTube, PDF, Word, audio, and video input formats and resource links.
• AI intelligent polishing: intelligently improves storytelling, clarity, and engagement, making podcasts more vivid and natural. Suitable for news, interviews, explanations, and various scenarios.
• Custom speech speed and style: supports customizable speech speed and style, multiple audio voices to meet diverse output needs.
• Perfect support for speakers and segments: supports multiple speakers, automatically distinguishes roles, easy segmentation of complex flows.
• High concurrency and fault tolerance: stable support for large-scale automated tasks, with status polling and clear error returns.
• Developer-friendly: open API, unified structured response, comprehensive documentation, easy for secondary integration.

Authentication

All API requests require api-key. Include your API key in request headers:

API Host: https://audiocleaner.ai/

api-key: YOUR_API_KEY

Overall Process (Two Steps)

• Call Step 1 create API to generate podcast script, return global_id
• Poll Step 1 status API to obtain the script result (train_data, podcast_script, etc.)
• Call Step 2 create API, input the script structure to generate the audio file, return global_id
• Poll Step 2 status API to get the final audio address (file_url)

1. Step 1 - Create Podcast Script Task

Endpoint

POST /audio/api/v1/api-keys/podcast/script/create

Request Headers

Request Parameters (Body · JSON)

Podcast Style Parameter Description

Duration Level Description

Supported Languages

Example Input (Plain Text)

{
  "source_type": 4,
  "speakers": [16098, 16093],
  "style": 1,
  "duration": 3,
  "use_mode": "ai",
  "commercial_use": 0,
  "language": "zh",
  "text": "Faced with growing backlash..."
}

Example Input (Attachment)

{
  "source_type": 3,
  "speakers": [16098, 16093],
  "style": 1,
  "duration": 3,
  "use_mode": "ai",
  "commercial_use": 0,
  "language": "zh",
  "url": "https://resource.audiocleaner.ai/paid/source-file/202604/14/1776147603_307691.jpg"
}

Example Input (Website)

{
  "source_type": 1,
  "speakers": [16098, 16093],
  "style": 1,
  "duration": 3,
  "use_mode": "ai",
  "commercial_use": 0,
  "language": "zh",
  "url": "https://audiocleaner.ai/"
}

Example Input (YouTube)

{
  "source_type": 2,
  "speakers": [16098, 16093],
  "style": 1,
  "duration": 3,
  "use_mode": "ai",
  "commercial_use": 0,
  "language": "zh",
  "url": "https://www.youtube.com/shorts/4QenPnBk99M"
}

Return Format (Success Example)

{
  "code": 100000,
  "data": {
    "global_id": "29a820e8-31ec-4502-859d-883be108d1e0"
  },
  "message": "Request Success"
}

Response Code (code) Description

Example

curl --location --request POST 'https://audiocleaner.ai/audio/api/v1/api-keys/podcast/script/create' \
  --header 'Content-Type: application/json' \
  --header 'api-key: YOUR_API_KEY' \
  --data-raw '{
    "source_type": 4,
    "speakers": [16098, 16093],
    "style": 1,
    "duration": 3,
    "use_mode": "ai",
    "commercial_use": 0,
    "language": "zh",
    "text": "Faced with growing backlash..."
  }'

2. Step 1 - Query Podcast Script Task Status

Endpoint

POST /audio/api/v1/api-keys/podcast/script/get

Request Headers

Request Parameters (Body · JSON)

Return Format (Success Example)

{
  "code": 100000,
  "data": {
    "status": "success",
    "global_id": "29a820e8-31ec-4502-859d-883be108d1e0",
    "source_data": "https://www.youtube.com/shorts/4QenPnBk99M",
    "train_data": "...",
    "podcast_script": [],
    "hightlights": [],
    "processed_content": "...",
    "title": "...",
    "avatar": null,
    "podcast_script_txt": "...",
    "create_time": 1776176539
  },
  "message": "Request Success"
}

Return Field Description

Response Code (code) Description

Example

curl --location --request POST 'https://audiocleaner.ai/audio/api/v1/api-keys/podcast/script/get' \
  --header 'Content-Type: application/json' \
  --header 'api-key: YOUR_API_KEY' \
  --data-raw '{
    "global_id": "29a820e8-31ec-4502-859d-883be108d1e0"
  }'

3. Step 2 - Create Podcast Audio Task

Endpoint

POST /audio/api/v1/api-keys/podcast/audio/create

Request Headers

Request Parameters (Body · JSON)

Return Format (Success Example)

{
  "code": 100000,
  "data": {
    "global_id": "29a820e8-31ec-4502-859d-883be108d1e0"
  },
  "message": "Request Success"
}

Response Code (code) Description

Example

curl --location --request POST 'https://audiocleaner.ai/audio/api/v1/api-keys/podcast/audio/create' \
  --header 'Content-Type: application/json' \
  --header 'api-key: YOUR_API_KEY' \
  --data-raw '{
    "title": "Who is Shelby? A Power Game in a Conversation",
    "processed_content": "In this episode, we explore the subtle moments of power encounters...",
    "podcast_script": [
      {
        "role": "A",
        "name": "Unrestrained Young Man",
        "avatar": "",
        "text": "Hi everyone! Today we’re going to talk about a very interesting conversation.",
        "voice_id": "https://resource.audiocleaner.ai/acweb/minimax/zh/Chinese (Mandarin)_Unrestrained_Young_Man.mp3"
      }
    ],
    "hightlights": [
      {
        "highlight": "Power encounter moments and provocative tone",
        "index": 0,
        "script_start_index": 0,
        "script_end_index": 3
      }
    ],
    "global_id": "ce2afb8f-7ccd-44a5-8bb0-8849098acfb7"
  }'

4. Step 2 - Query Podcast Audio Task Status

Endpoint

POST /audio/api/v1/api-keys/podcast/audio/get

Credit Consumption Description

• Credit consumption for audio generation tasks is billed according to the duration of the generated audio, as follows:
• Credit consumption = max(10, ⌊audio duration (minutes)⌋ × 10) (at least 10 credits)
• ⌊⌋ means rounding down. For example: 5.9 minutes is calculated as 5 minutes.
• If the current credit balance is less than required, the remaining balance will be set to zero (i.e., this task will not result in negative balance).

Request Headers

Request Parameters (Body · JSON)

Return Format (Success Example)

{
  "code": 100000,
  "data": {
    "status": "success",
    "global_id": "29a820e8-31ec-4502-859d-883be108d1e0",
    "file_url": "https://resource.audiocleaner.ai/paid/output/result/xxx_final.mp3",
    "source_data": "https://www.youtube.com/shorts/4QenPnBk99M",
    "train_data": "...",
    "podcast_script": [],
    "hightlights": [],
    "processed_content": "...",
    "title": "...",
    "avatar": "/img/cover1.png",
    "podcast_script_txt": "...",
    "create_time": 1776147898
  },
  "message": "Request Success"
}

Return Field Description (Key)

Response Code (code) Description

Example

curl --location --request POST 'https://audiocleaner.ai/audio/api/v1/api-keys/podcast/audio/get' \
  --header 'Content-Type: application/json' \
  --header 'api-key: YOUR_API_KEY' \
  --data-raw '{
    "global_id": "29a820e8-31ec-4502-859d-883be108d1e0"
  }'

5. Download Podcast Voice Resources

• Only returns voice information visible/available to the current account.
• Returns in JSON format and includes all required fields for business use, such as image/audio URLs and basic attributes (e.g. gender, age, language, etc).
• Key fields in the returned content are described, such as gender and age.
• After logging in, click the "Developer" menu, enter Playground, and select the "TAI Podcast Master" feature.

Return Format (Success Example)

{
  "code": 100000,
  "data": {
    "total": 481,
    "records": [
      {
        "id": 17827,
        "gender": 1,
        "gender_text": "male",
        "language": "en",
        "age": 2,
        "age_text": "young-adult",
        "emotion": [],
        "image": "https://resource.audiocleaner.ai/acweb/aicovergenerator/images/spongebob.webp",
        "audio": "https://resource.audiocleaner.ai/acweb/aicovergenerator/speak/spongebob.mp3",
        "voice_id": "",
        "tags": ["Animation", "Funny", "SpongeBob SquarePants"],
        "name": "spongebob",
        "display_name": "SpongeBob"
      }
      // ... more voice resources
    ]
  },
  "message": "Request Success"
}

Response Code (code) Description

Usage Flow Suggestions

• Call Step 1 create API according to the input source to get global_id
• Poll Step 1 query API until status=success and get podcast_script
• Use the script result as parameters to call Step 2 create API, get a new global_id
• Poll Step 2 query API until status=success and get file_url
• If resource or account-related data is needed, call the Podcast Voice Resource API

FAQ