Skip to main content

​
Overview

The Analyze API extracts meaningful insights, summaries, and custom information from video and audio content across major social media platforms. Simply provide a media URL and a question about the content, and the API will analyze the visual elements (if applicable) and transcriptions to deliver key takeaways and structured insights.

​
Features

  • Markdown Format: Receive analysis in formatted markdown, ideal for LLM applications.
  • Video Summaries: Generate concise summaries of video content for quick insights.
  • Custom Queries: Ask specific questions about the content to extract relevant information.
  • Brand Detection: Automatically identify brands displayed or mentioned in the content.
  • Full Transcription: Includes transcription and all its features for deeper analysis.
  • Platform Metadata: Optionally retrieve media metrics such as upload time, like count, comment count,and thumbnail URL.
  • Webhook Support: Get real-time notifications when the analysis is complete.

​
Basic Usage

To analyze content, simply add the media URL to the post_url parameter. You can optionally include a query to ask specific questions about the content, and a callback_url to receive notifications when the analysis is complete. 
curl --request POST \
  --url https://api.scribesocial.ai/v1/analyze \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '{
  "post_url": "<string>",
  "callback_url": "<string>",
  "query": null
}'

​
Checking Analysis Status (Async Only)

If you’re using a callback_url, the API response will include a Location header containing the operation polling URL, which you can use to check the status of the analysis. Example Operation Polling URL: https://api.scribesocial.ai/v1/transcribe-result/{operation-id} When the callback request is sent to your endpoint, it will include a Scribe-Verification-Token header. This token corresponds to the API key ID used for the original request, allowing you to verify that the callback is coming from dScribe AI and is legitimate.

​
Example Response

{
  "status": "completed",
  "operation_id": "abc123xyz789",
  "data": [
    {
      "video_summary": "In this video, the speaker provides a comprehensive introduction to machine learning fundamentals, covering basic concepts and practical applications in the healthcare industry...",
      "query_result": {
        "answer": "The main topics discussed are: 1) Introduction to ML concepts, 2) Healthcare applications, 3) Implementation challenges, 4) Future trends"
      },
      "detected_brands": [
        {
          "name": "TensorFlow",
          "description": "An open-source machine learning framework"
        },
        {
          "name": "PyTorch",
          "description": "Deep learning framework developed by Meta AI"
        }
      ],
      "video_url": "https://www.youtube.com/watch?v=example1",
      "post_url": "https://www.youtube.com/watch?v=example1",
      "description": "Part 1 of the machine learning tutorial series",
      "transcription": "Welcome to part one of our machine learning series...",
      "paragraphs": [
        {
          "sentences": [
            {
              "text": "Welcome to part one of our machine learning series.",
              "start": 0,
              "end": 3.2
            },
            {
              "text": "Today we'll cover the fundamental concepts.",
              "start": 3.2,
              "end": 6.5
            }
          ],
          "num_words": 18,
          "start": 0,
          "end": 6.5
        }
      ],
      "video_id": "example1",
      "transcribed_duration": 180,
      "total_duration": 180,
      "status": "completed",
      "platform": "youtube",
      "title": "Machine Learning Fundamentals - Part 1",
      "metadata": {
        "upload_time": "2024-01-15T10:30:00Z",
        "comment_count": 423,
        "like_count": 892,
        "thumbnail": "https://i.ytimg.com/vi/example1/maxresdefault.jpg",
        "retweet_count": null,
        "quote_count": null
      }
    }
  ],
  "markdown": "# Machine Learning Fundamentals - Part 1\n\n## Introduction (00:00)\nWelcome to part one of our machine learning series...\n\n## Key Concepts (01:23)\nLet's discuss the fundamental principles..."
}
For more details about request parameters, response fields, and status codes, check out our Analyze API Reference.

​
Common Use Cases

  1. LLM Applications: Enrich LLMs with structured visual insights and transcriptions to improve contextual understanding, retrieval, and summarization.
  2. Social Listening & Brand Monitoring: Track brand mentions, logos, and product placements in both video frames and spoken dialogue, enabling deeper sentiment and trend analysis.
  3. Content Intelligence & Analysis: Extract structured insights from visual elements and spoken words, categorizing information for better decision-making.
  4. Content Indexing & Searchability: Make video and audio content searchable by generating transcriptions and tagging visual elements such as objects, faces, and text overlays.
  5. Compliance & Moderation: Ensure compliance by analyzing both visual and audio content for policy violations, explicit material, and copyright concerns.
  6. Custom Queries & Deep Video Understanding: Ask specific questions about both the visual and spoken content to extract targeted insights, such as identifying key moments, detecting logos, or summarizing actions.