# LongCat-Video t2v

> Turn plain text into cinematic, on-brand video. Describe the scene and camera feel; LongCat generates smooth, consistent shots with adjustable length, FPS, and quality.

## Overview

- **Endpoint**: `https://queue.modelrunner.run/meituan-longcat/longcat-video-t2v-480p`
- **Model ID**: `meituan-longcat/longcat-video-t2v-480p`
- **Category**: text-to-video
- **Kind**: inference
- **Tags**: text-to-video

## Pricing

- **Price**: $0.02 per output second

## Request Lifecycle

This model runs on the ModelRunner **asynchronous queue API** — a single POST does not return the output. Every call requires an `Authorization: Key $MODEL_RUNNER_KEY` header. Run three steps:

1. **Submit** — `POST https://queue.modelrunner.run/meituan-longcat/longcat-video-t2v-480p` with a JSON body holding the input fields at the top level. The body may also include a reserved top-level `metadata` object — a flat string map (max 16 keys, key ≤64 / value ≤512 chars) stored on the request for your own tagging. It is never sent to the model; filter your request history with `GET https://queue.modelrunner.run/requests?metadata=<url-encoded JSON>` (exact key=value matches, AND-ed). The response carries request handles only (no output yet):

```json
{
  "status": "IN_QUEUE",
  "request_id": "<21-char id>",
  "status_url": "https://queue.modelrunner.run/meituan-longcat/longcat-video-t2v-480p/requests/<request_id>/status",
  "response_url": "https://queue.modelrunner.run/meituan-longcat/longcat-video-t2v-480p/requests/<request_id>",
  "cancel_url": "https://queue.modelrunner.run/meituan-longcat/longcat-video-t2v-480p/requests/<request_id>/cancel"
}
```

2. **Poll status** — `GET <status_url>` until `status` is `COMPLETED`. Possible values are `IN_QUEUE`, `IN_PROGRESS`, `COMPLETED`, `FAILED`, `CANCELLED`. A `FAILED` request responds with HTTP 400 and an `error` field.

3. **Read result** — `GET <response_url>`. Returns the finished request, including the generated `output`:

```json
{ "id": "<request_id>", "status": "COMPLETED", "output": ..., "input": ... }
```

The JavaScript and Python SDKs below perform steps 2–3 for you. In any language without an SDK (Swift, Go, Kotlin, etc.) you must implement the polling loop and the final result fetch yourself — see the cURL example for the full flow.

### Input Schema

- **`fps`** (`integer`, _optional_):
  The frame rate of the generated video.
  - Default: `15`
  - Range: `1` to `60`

- **`seed`** (`integer`, _optional_):
  The seed for the random number generator.
  - Default: `0`

- **`prompt`** (`string`, _required_):
  The prompt to guide the video generation.

- **`num_frames`** (`integer`, _optional_):
  The number of frames to generate.
  - Default: `162`
  - Range: `17` to `961`

- **`aspect_ratio`** (`aspect_ratio`, _optional_):
  The aspect ratio of the generated video.
  - Default: `"16:9"`
  - Options: `"16:9"`, `"9:16"`, `"1:1"`

- **`video_quality`** (`video_quality`, _optional_):
  The quality of the generated video.
  - Default: `"high"`
  - Options: `"low"`, `"medium"`, `"high"`, `"maximum"`

- **`video_write_mode`** (`video_write_mode`, _optional_):
  The write mode of the generated video.
  - Default: `"balanced"`
  - Options: `"fast"`, `"balanced"`, `"small"`

- **`video_output_type`** (`video_output_type`, _optional_):
  The output type of the generated video.
  - Default: `"X264 (.mp4)"`
  - Options: `"X264 (.mp4)"`, `"VP9 (.webm)"`, `"PRORES4444 (.mov)"`, `"GIF (.gif)"`

- **`num_inference_steps`** (`integer`, _optional_):
  The number of inference steps to use.
  - Default: `12`
  - Range: `2` to `16`

- **`enable_safety_checker`** (`boolean`, _optional_):
  Whether to enable safety checker. Can't be disabled in web UI.
  - Default: `true`

### Output Schema

_No `Output` schema properties are available._

## Default Example

**Input**

```json
{
  "fps": 15,
  "seed": 0,
  "prompt": "Realistic, cinematic filming style. An advanced skateboarder in a dark full-face helmet, deep-colored windbreaker, blue jeans, and bright yellow shoes rides along a winding alpine mountain road. Setting: lush green hills on both sides, distant snow-capped peaks, clear bright blue sky, crisp natural daylight. Shot duration ~8–10 seconds, single continuous take, no cuts.\n\nAction beats: at start the rider is already rolling; camera tracks closely from behind at low height (~1.5 m). The rider accelerates into an S-curve, then gradually lowers into a crouch; at mid-shot they extend the right hand to lightly touch the asphalt while keeping a low center of gravity to carve a sharp left hairpin. After completing the turn, the rider rises smoothly back to a standing position and continues gliding forward with relaxed balance.\n\nCamera direction: stabilized follow shot on a gimbal/drone/pursuit rig ~2–3 m behind, slight parallax to the road edge; gentle ease-in/ease-out speed changes; subtle lateral drift to reveal scenery; horizon level (no Dutch angles). Lens ~35 mm equivalent, medium depth of field; natural motion blur at 24–30 fps. Keep framing center-weighted with road lines leading forward; maintain ~60% ground, ~40% sky.\n\nLighting & mood: bright natural daylight with soft bounce; clean contrast, saturated greens, blue sky, white peak highlights; realistic color and textures. Physical realism: board trucks tilt into the turn, wheels grip the road, clothing and shoelaces react to wind, slight dust wisp where the hand grazes the surface.\n\nAvoid: jerky camera moves, jitter, rubbery or warped limbs, melting board/trucks, exaggerated speed ramps, floating shadows, text or watermarks. Keep everything coherent, grounded, and on-brand, delivering a smooth, premium commercial feel.",
  "num_frames": 162,
  "aspect_ratio": "16:9",
  "video_quality": "high",
  "video_write_mode": "balanced",
  "video_output_type": "X264 (.mp4)",
  "num_inference_steps": 15,
  "enable_safety_checker": false
}
```

**Output**

```json
"https://media.modelrunner.ai/zRoAzFbTjpDMmdP2pTZn5_tmpz44mxh1i.mp4"
```
## Usage Examples

### cURL

The queue API is asynchronous: submit the request, poll `status_url` until it is `COMPLETED`, then read the result from `response_url`. Requires `jq`.

```bash
# 1. Submit the request (returns request handles, not the output)
SUBMIT=$(curl --silent --request POST \
  --url https://queue.modelrunner.run/meituan-longcat/longcat-video-t2v-480p \
  --header "Authorization: Key $MODEL_RUNNER_KEY" \
  --header "Content-Type: application/json" \
  --data '{
  "fps": 15,
  "seed": 0,
  "prompt": "Realistic, cinematic filming style. An advanced skateboarder in a dark full-face helmet, deep-colored windbreaker, blue jeans, and bright yellow shoes rides along a winding alpine mountain road. Setting: lush green hills on both sides, distant snow-capped peaks, clear bright blue sky, crisp natural daylight. Shot duration ~8–10 seconds, single continuous take, no cuts.\n\nAction beats: at start the rider is already rolling; camera tracks closely from behind at low height (~1.5 m). The rider accelerates into an S-curve, then gradually lowers into a crouch; at mid-shot they extend the right hand to lightly touch the asphalt while keeping a low center of gravity to carve a sharp left hairpin. After completing the turn, the rider rises smoothly back to a standing position and continues gliding forward with relaxed balance.\n\nCamera direction: stabilized follow shot on a gimbal/drone/pursuit rig ~2–3 m behind, slight parallax to the road edge; gentle ease-in/ease-out speed changes; subtle lateral drift to reveal scenery; horizon level (no Dutch angles). Lens ~35 mm equivalent, medium depth of field; natural motion blur at 24–30 fps. Keep framing center-weighted with road lines leading forward; maintain ~60% ground, ~40% sky.\n\nLighting & mood: bright natural daylight with soft bounce; clean contrast, saturated greens, blue sky, white peak highlights; realistic color and textures. Physical realism: board trucks tilt into the turn, wheels grip the road, clothing and shoelaces react to wind, slight dust wisp where the hand grazes the surface.\n\nAvoid: jerky camera moves, jitter, rubbery or warped limbs, melting board/trucks, exaggerated speed ramps, floating shadows, text or watermarks. Keep everything coherent, grounded, and on-brand, delivering a smooth, premium commercial feel.",
  "num_frames": 162,
  "aspect_ratio": "16:9",
  "video_quality": "high",
  "video_write_mode": "balanced",
  "video_output_type": "X264 (.mp4)",
  "num_inference_steps": 15,
  "enable_safety_checker": false
}')

STATUS_URL=$(echo "$SUBMIT" | jq -r '.status_url')
RESPONSE_URL=$(echo "$SUBMIT" | jq -r '.response_url')

# 2. Poll until the request leaves the queue / in-progress state
while true; do
  STATUS=$(curl --silent --url "$STATUS_URL" \
    --header "Authorization: Key $MODEL_RUNNER_KEY" | jq -r '.status')
  echo "Status: $STATUS"
  case "$STATUS" in
    COMPLETED) break ;;
    FAILED|CANCELLED) echo "Request $STATUS"; exit 1 ;;
  esac
  sleep 1
done

# 3. Read the finished request, including the generated output
curl --silent --url "$RESPONSE_URL" \
  --header "Authorization: Key $MODEL_RUNNER_KEY"
```

### JavaScript

```javascript
import { modelrunner } from "@modelrunner/client";

const result = await modelrunner.subscribe("meituan-longcat/longcat-video-t2v-480p", {
  input: {
  "fps": 15,
  "seed": 0,
  "prompt": "Realistic, cinematic filming style. An advanced skateboarder in a dark full-face helmet, deep-colored windbreaker, blue jeans, and bright yellow shoes rides along a winding alpine mountain road. Setting: lush green hills on both sides, distant snow-capped peaks, clear bright blue sky, crisp natural daylight. Shot duration ~8–10 seconds, single continuous take, no cuts.\n\nAction beats: at start the rider is already rolling; camera tracks closely from behind at low height (~1.5 m). The rider accelerates into an S-curve, then gradually lowers into a crouch; at mid-shot they extend the right hand to lightly touch the asphalt while keeping a low center of gravity to carve a sharp left hairpin. After completing the turn, the rider rises smoothly back to a standing position and continues gliding forward with relaxed balance.\n\nCamera direction: stabilized follow shot on a gimbal/drone/pursuit rig ~2–3 m behind, slight parallax to the road edge; gentle ease-in/ease-out speed changes; subtle lateral drift to reveal scenery; horizon level (no Dutch angles). Lens ~35 mm equivalent, medium depth of field; natural motion blur at 24–30 fps. Keep framing center-weighted with road lines leading forward; maintain ~60% ground, ~40% sky.\n\nLighting & mood: bright natural daylight with soft bounce; clean contrast, saturated greens, blue sky, white peak highlights; realistic color and textures. Physical realism: board trucks tilt into the turn, wheels grip the road, clothing and shoelaces react to wind, slight dust wisp where the hand grazes the surface.\n\nAvoid: jerky camera moves, jitter, rubbery or warped limbs, melting board/trucks, exaggerated speed ramps, floating shadows, text or watermarks. Keep everything coherent, grounded, and on-brand, delivering a smooth, premium commercial feel.",
  "num_frames": 162,
  "aspect_ratio": "16:9",
  "video_quality": "high",
  "video_write_mode": "balanced",
  "video_output_type": "X264 (.mp4)",
  "num_inference_steps": 15,
  "enable_safety_checker": false
}
});

console.log(result.data);
```

### Python

```python
import asyncio
import modelrunner_ai

async def main():
    response = await modelrunner_ai.submit_async(
        "meituan-longcat/longcat-video-t2v-480p",
        arguments={
  "fps": 15,
  "seed": 0,
  "prompt": "Realistic, cinematic filming style. An advanced skateboarder in a dark full-face helmet, deep-colored windbreaker, blue jeans, and bright yellow shoes rides along a winding alpine mountain road. Setting: lush green hills on both sides, distant snow-capped peaks, clear bright blue sky, crisp natural daylight. Shot duration ~8–10 seconds, single continuous take, no cuts.\n\nAction beats: at start the rider is already rolling; camera tracks closely from behind at low height (~1.5 m). The rider accelerates into an S-curve, then gradually lowers into a crouch; at mid-shot they extend the right hand to lightly touch the asphalt while keeping a low center of gravity to carve a sharp left hairpin. After completing the turn, the rider rises smoothly back to a standing position and continues gliding forward with relaxed balance.\n\nCamera direction: stabilized follow shot on a gimbal/drone/pursuit rig ~2–3 m behind, slight parallax to the road edge; gentle ease-in/ease-out speed changes; subtle lateral drift to reveal scenery; horizon level (no Dutch angles). Lens ~35 mm equivalent, medium depth of field; natural motion blur at 24–30 fps. Keep framing center-weighted with road lines leading forward; maintain ~60% ground, ~40% sky.\n\nLighting & mood: bright natural daylight with soft bounce; clean contrast, saturated greens, blue sky, white peak highlights; realistic color and textures. Physical realism: board trucks tilt into the turn, wheels grip the road, clothing and shoelaces react to wind, slight dust wisp where the hand grazes the surface.\n\nAvoid: jerky camera moves, jitter, rubbery or warped limbs, melting board/trucks, exaggerated speed ramps, floating shadows, text or watermarks. Keep everything coherent, grounded, and on-brand, delivering a smooth, premium commercial feel.",
  "num_frames": 162,
  "aspect_ratio": "16:9",
  "video_quality": "high",
  "video_write_mode": "balanced",
  "video_output_type": "X264 (.mp4)",
  "num_inference_steps": 15,
  "enable_safety_checker": false
}
    )
    result = await response.get()
    print(result["output"])

asyncio.run(main())
```

## Additional Resources

- [Playground](https://modelrunner.ai/models/meituan-longcat/longcat-video-t2v-480p)
- [OpenAPI Schema](https://modelrunner.ai/models/meituan-longcat/longcat-video-t2v-480p/openapi.json)
- [LLM Instructions](https://modelrunner.ai/models/meituan-longcat/longcat-video-t2v-480p/llms.txt)
- [GitHub](https://github.com/meituan-longcat/LongCat-Video)
- [License](https://github.com/meituan-longcat/LongCat-Video/blob/main/LICENSE)
- [Paper](https://arxiv.org/abs/2510.22200)