# Asynchronous Inference

Asynchronous inference is the recommended way to call media models. You submit a request to a persistent queue, then retrieve results later by polling for status. 

The async approach gives you full control over the request lifecycle. You can submit many requests in parallel and process them as they complete, get real-time visibility into queue position and runner logs, and rely on automatic retries when failures occur. 

### Submit a Request

Submit generation tasks to via API.

{% tabs %}
{% tab title="Python" %}

```python
import http.client
import json

conn = http.client.HTTPSConnection("image.onerouter.pro")
payload = json.dumps({
   "model": "bytedance/seedance",
   "content": [
      {
         "type": "text",
         "text": "A little kitten walks along the seaside."
      }
   ],
   "generate_audio": True,
   "tools": [
      {
         "type": "web_search"
      }
   ],
   "resolution": "720p",
   "ratio": "adaptive",
   "duration": 5,
   "watermark": False
})
headers = {
   'Content-Type': 'application/json'
}
conn.request("POST", "/v1/images/generations", payload, headers)
res = conn.getresponse()
data = res.read()
print(data.decode("utf-8"))
```

{% endtab %}

{% tab title="cURL" %}

```bash
curl --location --request POST 'https://image.onerouter.pro/v1/images/generations' \
--header 'Content-Type: application/json' \
--data-raw '{
    "model": "bytedance/seedance",
    "content": [
      {
        "type": "text",
        "text": "A little kitten walks along the seaside."
      }
    ],
    "generate_audio": true,
    "tools": [
      {
        "type": "web_search"
      }
    ],
    "resolution": "720p",
    "ratio": "adaptive",
    "duration": 5,
    "watermark": false
}'
```

{% endtab %}
{% endtabs %}

Response format

```json
{
        "code": 200,
        "message": "success",
        "data": {
            "task_id": "cgt-20260312231129-7db6s",
            "object": "video",
            "model": "bytedance/seedance",
            "status": "created|in_progress|processing|completed|failed",
            "urls": {
                "get": "https://video.onerouter.pro/v1/videos/tasks/cgt-20260312231129-7db6s"
            },
        "created_at": "2024-01-01 12:00:00.00"
}
```

### Get the Result

Retrieve results from completed tasks.

{% tabs %}
{% tab title="Python" %}

```python
import http.client

conn = http.client.HTTPSConnection("video.onerouter.pro")
payload = ''
headers = {
   'Authorization': 'YOUR-API-KEY'
}
conn.request("GET", "/v1/videos/tasks/{task_id}", payload, headers)
res = conn.getresponse()
data = res.read()
print(data.decode("utf-8"))
```

{% endtab %}

{% tab title="cURL" %}

```bash
curl --location --request GET 'https://video.onerouter.pro/v1/videos/tasks/{task_id}' \
--header 'Authorization: sk-3c6klPYlaBYG3cboXrEWQFuGZSDDu3kCcyBRjkl10Er8arRt'
```

{% endtab %}
{% endtabs %}

Response format

```json
{
  "code": 200,
  "message": "success",
  "data": {
    "task_id": "cgt-20260312231129-7db6s",
    "object": "video",
    "model": "bytedance/seedance",
    "status": "created|in_progress|processing|completed|failed",
    "fail_reason": "",
    "submit_time": 1773318919,
    "start_time": 1773318925,
    "finish_time": 1773319107,
    "outputs": [
      "https://storage.googleapis.com/infron_gcs/video%2Fvideo%2Fcgt-20260328120757-nw49t_69c65308.mp4"
    ],
    "created_at": "2024-01-01 12:00:00"
  }
}

```