> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.reka.ai/vision/video-qa/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.reka.ai/_mcp/server.

# Video Q&A

> Ask questions about your videos and get AI-powered answers

The Vision API provides powerful question-answering capabilities for your videos. Once a video has been indexed, you can ask natural language questions about its content and receive AI-powered answers.

> **Info**
>
> The Video Q\&A API is designed for videos **longer than 30 seconds**. It indexes your video for efficient retrieval and analysis across longer content.
>
> For videos **30 seconds or shorter**, use the [Chat API with multimodal inputs](/chat/chat-with-image-video-and-audio) instead — it accepts video directly in the conversation without requiring a separate indexing step.

## Prerequisites

Before using Video Q\&A, ensure your video has been successfully indexed:

1. **Upload a video** with `index=true`
2. **Check indexing status** - it should be `"indexed"`
3. **Wait for processing** if status is `"indexing"`

## Ask Questions

#### Bash

Use the `/v1/qa/chat` endpoint to ask questions about your videos:

```bash
curl -X POST https://vision-agent.api.reka.ai/v1/qa/chat \
  -H "X-Api-Key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "video_id": "550e8400-e29b-41d4-a716-446655440000",
    "messages": [
      {
        "role": "user",
        "content": "What is happening in this video?"
      }
    ]
  }'
```

#### Python

```python
import requests
import json

url = f"{BASE_URL}/v1/qa/chat"

# Chat request body
payload = {
    "video_id": "550e8400-e29b-41d4-a716-446655440000",
    "messages": [
      {
        "role": "user",
        "content": "What is happening in this video?"
      }
    ]
}

headers = {
    "X-Api-Key": REKA_API_KEY,
    "Content-Type": "application/json",
}

response = requests.post(url, json=payload, headers=headers)
print(response.status_code, response.json())
```

### Request Parameters

* **`video_id`** (optional): ID of the video to analyze
* **`messages`** (optional): List of chat messages

### Using Chat Messages

For multi-turn conversations, use the `messages` parameter:

#### Bash

```bash
curl -X POST https://vision-agent.api.reka.ai/v1/qa/chat \
  -H "X-Api-Key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "video_id": "550e8400-e29b-41d4-a716-446655440000",
    "messages": [
      {
        "role": "user",
        "content": "What is happening in this video?"
      },
      {
        "role": "assistant",
        "content": "The video shows a person walking on the beach during sunset."
      },
      {
        "role": "user",
        "content": "What is the color of the car in the video?"
      }
    ]
  }'
```

#### Python

```python
import requests
import json

url = f"{BASE_URL}/v1/qa/chat"

# Chat request body
payload = {
    "video_id": "550e8400-e29b-41d4-a716-446655440000",
    "messages": [
      {
        "role": "user",
        "content": "What is happening in this video?"
      },
      {
        "role": "assistant",
        "content": "The video shows a person walking on the beach during sunset."
      },
      {
        "role": "user",
        "content": "What is the color of the car in the video?"
      }
    ]
}

headers = {
    "X-Api-Key": REKA_API_KEY,
    "Content-Type": "application/json",
}

response = requests.post(url, json=payload, headers=headers)
print(response.status_code, response.json())
```

## Streaming Responses

For real-time responses, set `stream=true` in your request:

#### Bash

```bash
curl -X POST https://vision-agent.api.reka.ai/v1/qa/chat \
  -H "X-Api-Key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "video_id": "550e8400-e29b-41d4-a716-446655440000",
    "stream": true,
    "messages": [
      {
        "role": "user",
        "content": "What is happening in this video?"
      }
    ]
  }'
```

#### Python

```python
import requests
import json

url = f"{BASE_URL}/v1/qa/chat"

payload = {
    "video_id": "550e8400-e29b-41d4-a716-446655440000",
    "stream": True,
    "messages": [
      {
        "role": "user",
        "content": "What is happening in this video?"
      }
    ]
}

headers = {
    "X-Api-Key": REKA_API_KEY,
    "Content-Type": "application/json",
}

response = requests.post(url, json=payload, headers=headers, stream=True)
for line in response.iter_lines():
    if line:
        print(line.decode('utf-8'))
```

This returns a Server-Sent Events (SSE) stream with real-time updates.

## Response Format

### Chat Response

```json
{
  "chat_response": "The video shows a person walking on the beach during sunset.",
  "status": "success",
}
```

### Stream Response

```json
{
  "event": "qa_stream",
  "data": {
    "chat_response": "The video shows a person walking on the beach during sunset.",
    "status": "success",
  }
}
```

When the stream is complete, a final event is sent:

```json
{
  "event": "done",
  "data": "[DONE]"
}
```

## Question Examples

Here are some example questions you can ask:

* **General**: "What is happening in this video?"
* **Specific**: "What color is the car in the video?"
* **Temporal**: "What happens at the beginning of the video?"
* **Analytical**: "How many people are in the scene?"
* **Descriptive**: "Describe the setting and atmosphere"

## Best Practices

1. **Be specific** in your questions for better answers
2. **Check indexing status** before asking questions
3. **Use streaming** for long videos or complex questions

## Error Handling

* **Video not found**: Ensure the video\_id is correct
* **Video not indexed**: Wait for indexing to complete
* **Indexing failed**: Re-upload the video with `index=true`

> Documentation for Reka AI APIs.