Skip to content
Nappai Documentation

Describe Video

Describe a video using AI to generate a textual summary.

How it Works

The component takes a video file (or a link to one) and sends it to an AI model that can understand both the visual and audio parts of the clip. First, the video is broken into frames and its audio track is transcribed. Then the language model reads the transcript and the visual cues and creates a short description based on the prompt you give. The result is a list of data objects that you can use elsewhere in your workflow.

Inputs

  • Audio Transcriber: Choose an audio transcriber that will read the sound track of the video and turn it into text. This text is used by the language model to help describe what happens in the video.
  • Model: Pick the language model that will analyze the video and write the description. The model must be able to understand text and visual information.
  • Video Data: Provide the video you want described. You can drop a file, paste a URL, or give a list of video objects. The component will handle any of these formats.
  • Prompt Text: Write a short instruction for the model, such as “Describe the video.” This tells the AI what kind of description you want.

Outputs

  • Data: The component returns a list of Data objects. Each object contains the description text and any metadata the model produced. You can feed this output into other components, store it, or display it in the dashboard.

Usage Example

  1. Drag the Describe Video component onto the canvas.
  2. Connect a Video component (or upload a file) to the Video Data input.
  3. Select a Language Model (e.g., GPT‑4) for the Model input.
  4. Optionally choose an Audio Transcriber if you want the audio to be processed separately.
  5. Leave the Prompt Text as “Describe the video.”
  6. Run the workflow. The output will be a short description of the video that you can display or store.
  • Upload Video – Lets you add a video file to the workflow.
  • Text Summarizer – Summarizes long text; can be used after the video description is generated.
  • Store Data – Saves the description to a database or file.

Tips and Best Practices

  • Use a high‑quality audio transcriber for clearer descriptions.
  • Keep the prompt short and clear; “Describe the video” works well for most cases.
  • If the video is long, consider trimming it first to reduce processing time.
  • Test with a short clip before running the full workflow to ensure the model understands the content.

Security Considerations

  • The video and its transcript are sent to the chosen language model, which may be hosted by a third‑party provider. Make sure you trust the provider and that any sensitive data is handled according to your organization’s privacy policies.
  • If the video contains confidential information, consider using an on‑premises model or a secure, private API endpoint.