Video - PraisonAI

Agents can analyze videos - describe content, answer questions, and extract key moments.

Quick Start

Simple Usage

import { Agent } from 'praisonai';

const agent = new Agent({
  instructions: 'You analyze videos and describe what you see',
  video: true
});

await agent.chat('Analyze https://example.com/video.mp4');

With Configuration

await agent.chat([
  { role: 'user', content: [
    { type: 'text', text: 'What is happening in this video?' },
    { type: 'video', url: 'https://example.com/video.mp4' }
  ]}
]);

User Interaction Flow

Configuration Levels

Based on your needs, configure video analysis at different levels:

// Level 1: Bool - Enable with defaults
const agent = new Agent({ video: true });

// Level 2: Dict - Custom options
const agent = new Agent({
  video: {
    maxDuration: 300,
    frameRate: 1
  }
});

// Level 3: Instance - Full control
import { VideoAgent } from 'praisonai';

const videoAgent = new VideoAgent({
  llm: 'gpt-4o',
  maxDuration: 600,
  transcribeAudio: true
});

API Reference

VideoConfig

Complete configuration options

VideoAgent

Full class documentation

What You Can Do

Action	Example
Describe video	`agent.chat('What happens in this video?')`
Ask questions	`agent.chat('How many people appear?')`
Get summary	`agent.chat('Summarize this in 2 sentences')`
Find moments	`agent.chat('List the key events')`

Best Practices

Use GPT-4o for video

GPT-4o has the best video understanding. Other models may have limited support.

Keep videos short

Videos under 5 minutes work best. Split longer videos for better analysis.

Be specific in questions

Ask about specific moments or details rather than general questions.

Vision

Analyze images

Audio

Process audio content

​Quick Start

​User Interaction Flow

​Configuration Levels

​API Reference