Skip to main content

Describe Video Node

The Describe Video node uses AI to analyze video content and generate descriptions, summaries, or insights based on custom prompts. This powerful node leverages advanced AI models to extract meaningful information from videos, including identifying topics, subjects, actions, scenes, and more, making video content searchable and accessible.

Describe Video node

Basic Usage​

Use the Video Input, Text Node, Describe Video, and Display Text nodes to create a video analysis workflow.

Inputs​

The Describe Video node accepts the following inputs:

  • Input: The video file or video URL to be analyzed by the AI model.
  • Reference: Additional context or reference information to guide the video analysis (optional).
  • Overwrite System Prompt: Custom instructions that completely replace the default AI behavior for video analysis (optional).

Outputs​

  • Output: The AI-generated description, analysis, or summary of the video content based on the input and system instructions.

Configuration​

AI Model Selection​

Select the AI model to use for video analysis from the dropdown menu:

  • Gemini 2.5 Flash Lite: Fast and cost-effective option for most video analysis tasks.
  • Gemini 1.5 Pro: Advanced model with enhanced understanding capabilities.
  • Gemini 1.5 Flash: Balanced performance for general video description tasks.
  • Other supported models: Additional AI models available in your system configuration.

Note: Video analysis capabilities depend on the selected model's multimodal support.

Example Workflows​

Video Content Analysis​

Scenario: Analyze an uploaded video to identify the main topic and key individuals or subjects discussed.

Describe Video Example

Steps to Create the Flow:

  1. Start with the Start Node.

  2. Add and connect a Video Input node for users to upload videos.

  3. Add a Text node with analysis instructions:

    i. Create a Text node with your analysis prompt

    • Example prompt:
    You are a professional content curator. Analyze the video to identify the **main topic**, 
    **key individuals/subjects**, and provide a brief summary of the content.

    ii. Configure text options:

    • Optionally check "Encrypt the text" if handling sensitive instructions

  4. Add and connect a Describe Video node:

    i. Connect the Video Input's Output to Input:

    • Connect the video file from the Video Input node to the Input handle

    ii. Connect the Text node's Output to Overwrite System Prompt:

    • This provides custom instructions for the AI analysis

    iii. Select the AI Model:

    • Choose Gemini 2.5 Flash Lite (or your preferred model from the dropdown)

  5. Add Display Text to show the analysis results.

Preview:

Video Content:

  • A 5-minute educational video about climate change featuring Dr. Jane Smith

AI Analysis Output:

**Main Topic:** Climate Change and Global Warming

**Key Individuals/Subjects:**
- Dr. Jane Smith - Environmental Scientist and Presenter
- Discussion of greenhouse gases
- Impact on polar ice caps
- Renewable energy solutions

**Summary:**
Dr. Jane Smith presents an overview of climate change, explaining the causes of global 
warming through greenhouse gas emissions. The video covers the observable impacts on 
polar regions and discusses potential solutions through renewable energy adoption. 
The presentation uses visual aids and data charts to illustrate temperature trends 
over the past century.

Result: Detailed video analysis that makes content discoverable and provides structured insights into video content.

  • Video Input: Upload or capture video files for analysis
  • Display Text: Show the video analysis results
  • AI General Prompt: Process analysis output further
  • Describe Image: Analyze individual video frames or screenshots
  • JSON Builder: Structure video analysis into JSON format
  • Data Dump: Store video analysis results
  • Branching Node: Route based on video analysis outcomes
  • Online Video Player: Display the video alongside analysis
  • Transcribe Speech: Extract audio transcript from video
  • Form Node: Collect additional context for video analysis