---
title: Video Inputs
description: Send video files to video-capable models through the Agnic AI Gateway
---

# Video Inputs

Send video files to compatible models for analysis, description, object detection, and action recognition. Agnic supports multiple video formats for comprehensive video understanding tasks.

## Overview

Video requests are available via the `/v1/chat/completions` API with the `video_url` content type. Videos can be sent as URLs or base64-encoded data URLs.

<Callout type="info">
  Video URL support varies by provider. Some providers only support specific URL types (e.g., YouTube links for certain Google models).
</Callout>

---

## Using Video URLs

For publicly accessible videos:

### Python

```python
from openai import OpenAI

client = OpenAI(
    api_key="agnic_tok_YOUR_TOKEN",
    base_url="https://api.agnic.ai/v1"
)

response = client.chat.completions.create(
    model="google/gemini-2.0-flash",
    messages=[
        {
            "role": "user",
            "content": [
                {
                    "type": "text",
                    "text": "Please describe what's happening in this video."
                },
                {
                    "type": "video_url",
                    "video_url": {
                        "url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ"
                    }
                }
            ]
        }
    ]
)

print(response.choices[0].message.content)
```

### JavaScript

```javascript
import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: 'agnic_tok_YOUR_TOKEN',
  baseURL: 'https://api.agnic.ai/v1'
});

const response = await client.chat.completions.create({
  model: 'google/gemini-2.0-flash',
  messages: [
    {
      role: 'user',
      content: [
        {
          type: 'text',
          text: "Please describe what's happening in this video."
        },
        {
          type: 'video_url',
          video_url: {
            url: 'https://www.youtube.com/watch?v=dQw4w9WgXcQ'
          }
        }
      ]
    }
  ]
});

console.log(response.choices[0].message.content);
```

---

## Using Base64 Encoded Videos

For local video files:

### Python

```python
from openai import OpenAI
import base64

def encode_video(video_path):
    with open(video_path, "rb") as video_file:
        return base64.b64encode(video_file.read()).decode('utf-8')

client = OpenAI(
    api_key="agnic_tok_YOUR_TOKEN",
    base_url="https://api.agnic.ai/v1"
)

# Read and encode the video
video_path = "path/to/your/video.mp4"
base64_video = encode_video(video_path)
data_url = f"data:video/mp4;base64,{base64_video}"

response = client.chat.completions.create(
    model="google/gemini-2.0-flash",
    messages=[
        {
            "role": "user",
            "content": [
                {
                    "type": "text",
                    "text": "What's happening in this video?"
                },
                {
                    "type": "video_url",
                    "video_url": {
                        "url": data_url
                    }
                }
            ]
        }
    ]
)

print(response.choices[0].message.content)
```

### JavaScript

```javascript
import OpenAI from 'openai';
import fs from 'fs';

const client = new OpenAI({
  apiKey: 'agnic_tok_YOUR_TOKEN',
  baseURL: 'https://api.agnic.ai/v1'
});

// Read and encode the video
const videoPath = 'path/to/your/video.mp4';
const videoBuffer = fs.readFileSync(videoPath);
const base64Video = videoBuffer.toString('base64');
const dataUrl = `data:video/mp4;base64,${base64Video}`;

const response = await client.chat.completions.create({
  model: 'google/gemini-2.0-flash',
  messages: [
    {
      role: 'user',
      content: [
        {
          type: 'text',
          text: "What's happening in this video?"
        },
        {
          type: 'video_url',
          video_url: {
            url: dataUrl
          }
        }
      ]
    }
  ]
});

console.log(response.choices[0].message.content);
```

---

## Supported Video Formats

| Format | MIME Type | Extension |
|--------|-----------|-----------|
| MP4 | `video/mp4` | `.mp4` |
| MPEG | `video/mpeg` | `.mpeg` |
| MOV | `video/mov` | `.mov` |
| WebM | `video/webm` | `.webm` |

---

## Compatible Models

Models with video processing capabilities:

| Provider | Models | URL Support |
|----------|--------|-------------|
| Google | Gemini 2.0 Flash, Gemini 1.5 Pro | YouTube links |

---

## Use Cases

- **Video summarization** - Generate text summaries
- **Object recognition** - Identify objects and people
- **Action detection** - Describe activities and movements
- **Scene understanding** - Analyze settings and contexts
- **Content moderation** - Review video content
- **Sports analysis** - Analyze gameplay and movements

---

## Best Practices

### File Size Considerations

- **Compress videos** - Reduce size without losing quality
- **Trim videos** - Include only relevant segments
- **Lower resolution** - 720p is often sufficient
- **Reduce frame rate** - Lower FPS for non-critical analysis

### Optimal Video Length

- Check model-specific limits on duration
- Split long videos into shorter segments
- Focus on key moments

### Quality vs. Size Trade-offs

| Quality | Resolution | Use Case |
|---------|------------|----------|
| High | 1080p+ | Detailed analysis, text recognition |
| Medium | 720p | General analysis |
| Lower | 480p | Basic scene understanding |

---

## Troubleshooting

**Video not processing?**
- Verify the model supports video input
- Check format compatibility
- Try base64 encoding instead of URL
- Ensure file isn't corrupted

**Large file errors?**
- Compress the video
- Reduce resolution or frame rate
- Trim to shorter duration

**Poor analysis?**
- Ensure video quality is sufficient
- Provide specific prompts
- Check if content is clearly visible

<Cards>
  <Card title="Audio Inputs" href="/docs/ai-gateway/multimodal/audio" />
  <Card title="Image Inputs" href="/docs/ai-gateway/multimodal/images" />
</Cards>