Media Service API Documentation

Endpoint

• Method: POST
• Path: https://api.kadal.ai/media-services/api/v1/
• Summary: Extracts transcription, description, metadata, and thumbnail from a media file.

Description

This endpoint processes a single audio or video file to generate multiple outputs, including transcription of speech, a descriptive summary, technical metadata, and a thumbnail image. It enables comprehensive analysis and representation of media content in a single request.

Request

• Content-Type: multipart/form-data

Payload

Parameter	Description	Data Type	Is Optional
choice_model_selection	AI model selection (whisper)	String	No
media_type	Select the type of uploaded file (audio/video)	String	No
input_media	Upload a single audio/video file	File (.mp3, .mp4, .mpeg, .wav, .webm)	No
description	Enable detailed description generation	boolean	Yes
description_length	Desired length of generated description (25-500)	integer	Yes
metadata	Enable metadata generation	boolean	Yes
thumbnail	Enable thumbnail generation (only supported for video files)	boolean	Yes

Response

{
  "status_code": 200,
  "message": "SUCCESS::MediaService generated successfully",
  "file_name": "Your filename will appear here",
  "transcription": "Transcription will appear here",
  "description_text": "Description will appear here",
  "metadata_output": {
    "title": "Title will appear here",
    "description": "Description will appear here",
    "genre": ["Genre will appear here"],
    "creator": "Creator will appear here",
    "keywords": ["Keywords will appear here"],
    "custom": "custom will appear here"
  },
  "thumbnail_output": "Thumbnail will appear here"
}

Usage

import requests

url = "https://api.kadal.ai/media-services/api/v1/"
file_path = "your file path here"
token = "your token here"

headers = {
    "accept": "application/json",
    "Authorization": f"Bearer {token}",
}

data = {
    "choice_model_selection": "whisper",
    "media_type": "audio",
    "description": "true",
    "description_length": "225",
    "metadata": "true",
    "thumbnail": "false",
}

files = {
    "input_media": (file_path, open(file_path, "rb"), "audio/wav"),
}

response = requests.post(url, headers=headers, data=data, files=files)