# Generating Videos

Source: https://docs.prodia.com/guides/generating-videos/

This guide uses Veo Fast for text-to-video and Seedance for image-to-video. You can swap the job type to use any of these video generation models:

- [Wan 2.2 Lightning](https://docs.prodia.com/models/wan-2-2/) — `inference.wan2-2.lightning.txt2vid.v0` and `img2vid` (fast, \~22s)
- [Wan 2.7](https://docs.prodia.com/models/wan-2-7/) — `inference.wan2-7.txt2vid.v1` and more (1080p, up to 15s, audio-driven)
- [Veo](https://docs.prodia.com/models/veo/) — `inference.veo.fast.txt2vid.v2` and more (with audio generation)
- [Kling](https://docs.prodia.com/models/kling/) — `inference.kling.txt2vid.v1` (camera control, motion masks)
- [Seedance](https://docs.prodia.com/models/seedance/) — `inference.seedance.proturbo.txt2vid.v1` and `img2vid` (1080p, \~45s, async-only)

### Project Setup

```bash
# Create a project directory.
mkdir prodia-video-generation
cd prodia-video-generation
```

### prodia-js

Install Node (if not already installed):

### macOS

```bash
brew install node
# Close the current terminal and open a new one so that node is available.
```

### Linux

```bash
apt install node
# Close the current terminal and open a new one so that node is available.
```

### Windows

```bash
winget install -e --id OpenJS.NodeJS.LTS
# Close the current terminal and open a new one so that node is available.
```

Create project skeleton:

```bash
# Requires node --version >= 18
# Initialize the project with npm.
npm init -y

# Install the prodia-js library.
npm install prodia --save
```

### requests

Install Python (if not already installed):

### macOS

```bash
brew install python
# Close the current terminal and open a new one so that python is available.
```

### Linux

```bash
apt install python3 python3-venv python-is-python3
# Close the current terminal and open a new one so that python is available.
```

### Windows

```bash
winget install -e --id Python.Python.3.12
# Close the current terminal and open a new one so that python is available.
```

```bash
# Requires python --version >= 3.12
python -m venv venv
source venv/bin/activate
pip install requests
```

### curl

Install curl (if not already installed):

### macOS

```bash
brew install curl
# Close the current terminal and open a new one so that curl is available.
```

### Linux

```bash
apt install curl
# Close the current terminal and open a new one so that curl is available.
```

### Windows

```bash
# NOTE: Windows 10 and up have curl installed by default and this can be
# skipped.
winget install -e --id cURL.cURL
# Close the current terminal and open a new one so that curl is available.
```

```bash
# Export your token so it can be used by the main code.
export PRODIA_TOKEN=your-token-here
```

> note:
>
> Your token is exported to an environment variable. If you close or switch your
> shell you'll need to run `export PRODIA_TOKEN=your-token-here` again.

Create a main file for your project:

### prodia-js

```js title="main.js"
const { createProdia } = require("prodia/v2");

const prodia = createProdia({
    token: process.env.PRODIA_TOKEN // get it from environment
});
```

### requests

Create the following `main.py`

```python title="main.py"
from requests.adapters import HTTPAdapter, Retry
import os
import requests
import sys


prodia_token = os.getenv('PRODIA_TOKEN')
prodia_url = 'https://inference.prodia.com/v2/job'

session = requests.Session()
retries = Retry(allowed_methods=None, status_forcelist=Retry.RETRY_AFTER_STATUS_CODES)
session.mount('http://', HTTPAdapter(max_retries=retries))
session.mount('https://', HTTPAdapter(max_retries=retries))
session.headers.update({'Authorization': f"Bearer {prodia_token}"})
```

### curl

```bash title="main.sh"
set -euo pipefail
```

You're now ready to make some API calls!

### With a Prompt

### prodia-js

```javascript
const { createProdia } = require("prodia/v2");
const fs = require("node:fs/promises"); // add this to imports at the top

const prodia = createProdia({
  token: process.env.PRODIA_TOKEN, // get it from environment
});

(async () => {
  const job = await prodia.job({
    type: "inference.veo.fast.txt2vid.v1",
    config: {
      prompt: "A sweeping mountain landscape at sunrise, captured from a high-angle perspective using a wide-angle lens. The early morning light casts long shadows across the rugged terrain, with mist rolling over the valleys. The scene features sharp detail in the rocks, lush greenery, and clouds forming over distant peaks. Warm oranges and pinks dominate the sky, creating a dramatic and serene atmosphere. High dynamic range (HDR) captures the subtle transitions between light and shadow.",
    },
  });

  const video = await job.arrayBuffer();

  await fs.writeFile("landscape.mp4", new Uint8Array(video));
})();
```

```bash
node main.js
```

### requests

```python
from requests.adapters import HTTPAdapter, Retry
import os
import requests
import sys


prodia_token = os.getenv('PRODIA_TOKEN')
prodia_url = 'https://inference.prodia.com/v2/job'

session = requests.Session()
retries = Retry(status_forcelist=Retry.RETRY_AFTER_STATUS_CODES)
session.mount('http://', HTTPAdapter(max_retries=retries))
session.mount('https://', HTTPAdapter(max_retries=retries))
session.headers.update({'Authorization': f"Bearer {prodia_token}"})

headers = {
    'Accept': 'video/mp4',
}

job = {
    'type': 'inference.veo.fast.txt2vid.v1',
    'config': {
        'prompt': 'A sweeping mountain landscape at sunrise, captured from a high-angle perspective using a wide-angle lens. The early morning light casts long shadows across the rugged terrain, with mist rolling over the valleys. The scene features sharp detail in the rocks, lush greenery, and clouds forming over distant peaks. Warm oranges and pinks dominate the sky, creating a dramatic and serene atmosphere. High dynamic range (HDR) captures the subtle transitions between light and shadow.',
    },
}

res = session.post(prodia_url, headers=headers, json=job)
print(f"Request ID: {res.headers['x-request-id']}")
print(f"Status: {res.status_code}")

if res.status_code != 200:
    print(res.text)
    sys.exit(1)

with open('landscape.mp4', 'wb') as f:
    f.write(res.content)
```

```bash
python main.py
```

### curl

```bash
set -euo pipefail

job=$(cat <<EOF
{
  "type": "inference.veo.fast.txt2vid.v1",
  "config": {
    "prompt": "A sweeping mountain landscape at sunrise, captured from a high-angle perspective using a wide-angle lens. The early morning light casts long shadows across the rugged terrain, with mist rolling over the valleys. The scene features sharp detail in the rocks, lush greenery, and clouds forming over distant peaks. Warm oranges and pinks dominate the sky, creating a dramatic and serene atmosphere. High dynamic range (HDR) captures the subtle transitions between light and shadow."
  }
}
EOF
)

curl -sSf \
  -H "Authorization: Bearer $PRODIA_TOKEN" \
  -H 'Accept: video/mp4' \
  --json "$job" \
  --output landscape.mp4 \
  --retry 3 \
  https://inference.prodia.com/v2/job
```

```bash
bash main.sh
```

### macOS

```bash
open landscape.mp4
```

### Linux

```bash
xdg-open landscape.mp4
```

### Windows

```bash
start landscape.mp4
```

### From an Image

Let's extend our example and generate a video from an existing image.

### prodia-js

```javascript
const { createProdia } = require("prodia/v2");
const fs = require("node:fs/promises"); // add this to imports at the top

const prodia = createProdia({
  token: process.env.PRODIA_TOKEN, // get it from environment
});

(async () => {
  // get input image
  const inputBuffer = await (await fetch("https://docs.prodia.com/strike-a-pose.jpg")).arrayBuffer();

  const job = await prodia.job({
    type: "inference.veo.fast.img2vid.v1",
    config: {
      prompt: "Walking down the street.",
    },
  }, {
    inputs: [ inputBuffer ]
  });

  const video = await job.arrayBuffer();

  await fs.writeFile("walking.mp4", new Uint8Array(video));
})();
```

```bash
node main.js
```

### requests

```python
from requests.adapters import HTTPAdapter, Retry
from io import BytesIO
import json
import os
import requests
import sys


prodia_token = os.getenv('PRODIA_TOKEN')
prodia_url = 'https://inference.prodia.com/v2/job'

session = requests.Session()
retries = Retry(status_forcelist=Retry.RETRY_AFTER_STATUS_CODES)
session.mount('http://', HTTPAdapter(max_retries=retries))
session.mount('https://', HTTPAdapter(max_retries=retries))
session.headers.update({'Authorization': f"Bearer {prodia_token}"})

try:
    with open('strike-a-pose.jpg', 'rb') as f:
        input_image = f.read()
except FileNotFoundError:
    res = requests.get('https://docs.prodia.com/strike-a-pose.jpg')
    input_image = BytesIO(res.content)
    with open('strike-a-pose.jpg', 'wb') as f:
        f.write(res.content)
except Exception as e:
    raise e

headers = {
    'Accept': 'video/mp4',
}

job = {
    'type': 'inference.veo.fast.img2vid.v1',
    'config': {
        'prompt': 'Walking down the street.',
    },
}

files = [
    ('job', ('job.json', BytesIO(json.dumps(job).encode('utf-8')), 'application/json')),
    ('input', ('strike-a-pose.jpg', input_image, 'image/jpeg')),
]

res = session.post(prodia_url, headers=headers, files=files)
print(f"Request ID: {res.headers['x-request-id']}")
print(f"Status: {res.status_code}")

if res.status_code != 200:
    print(res.text)
    sys.exit(1)

with open('walking.mp4', 'wb') as f:
    f.write(res.content)
```

```bash
python main.py
```

### curl

```bash
set -euo pipefail

cat <<EOF > job.json
{
  "type": "inference.veo.fast.img2vid.v1",
  "config": {
    "prompt": "Walking down the street."
  }
}
EOF

if [[ ! -f strike-a-pose.jpg ]]; then
  curl -Lo strike-a-pose.jpg 'https://docs.prodia.com/strike-a-pose.jpg'
fi

curl -sSf \
  -H "Authorization: Bearer $PRODIA_TOKEN" \
  -H 'Accept: video/mp4' \
  -F job=@job.json \
  -F input=@strike-a-pose.jpg \
  --output walking.mp4 \
  --retry 3 \
  https://inference.prodia.com/v2/job
```

```bash
bash main.sh
```

### macOS

```bash
open walking.mp4
```

### Linux

```bash
xdg-open walking.mp4
```

### Windows

```bash
start walking.mp4
```

### With Audio

Veo can generate a synchronized audio track alongside the video — water, footsteps,
ambient sound, dialogue with lip sync — in a single job. Set `generate_audio: true`
and describe the soundscape in your prompt.

### prodia-js

```javascript
const { createProdia } = require("prodia/v2");
const fs = require("node:fs/promises"); // add this to imports at the top

const prodia = createProdia({
  token: process.env.PRODIA_TOKEN, // get it from environment
});

(async () => {
  const job = await prodia.job({
    type: "inference.veo.fast.txt2vid.v2",
    config: {
      prompt: "A wooden water mill in a forest stream, water splashing on the wheel, leaves rustling, peaceful afternoon, cinematic HDR",
      negative_prompt: "low quality, blurry, watermark",
      resolution: "720p",
      aspect_ratio: "16:9",
      duration_seconds: 4,
      generate_audio: true,
    },
  });

  const video = await job.arrayBuffer();

  await fs.writeFile("watermill.mp4", new Uint8Array(video));
})();
```

```bash
node main.js
```

### requests

```python
from requests.adapters import HTTPAdapter, Retry
import os
import requests
import sys


prodia_token = os.getenv('PRODIA_TOKEN')
prodia_url = 'https://inference.prodia.com/v2/job'

session = requests.Session()
retries = Retry(status_forcelist=Retry.RETRY_AFTER_STATUS_CODES)
session.mount('http://', HTTPAdapter(max_retries=retries))
session.mount('https://', HTTPAdapter(max_retries=retries))
session.headers.update({'Authorization': f"Bearer {prodia_token}"})

headers = {
    'Accept': 'video/mp4',
}

job = {
    'type': 'inference.veo.fast.txt2vid.v2',
    'config': {
        'prompt': 'A wooden water mill in a forest stream, water splashing on the wheel, leaves rustling, peaceful afternoon, cinematic HDR',
        'negative_prompt': 'low quality, blurry, watermark',
        'resolution': '720p',
        'aspect_ratio': '16:9',
        'duration_seconds': 4,
        'generate_audio': True,
    },
}

res = session.post(prodia_url, headers=headers, json=job)
print(f"Request ID: {res.headers['x-request-id']}")
print(f"Status: {res.status_code}")

if res.status_code != 200:
    print(res.text)
    sys.exit(1)

with open('watermill.mp4', 'wb') as f:
    f.write(res.content)
```

```bash
python main.py
```

### curl

```bash
set -euo pipefail

job=$(cat <<EOF
{
  "type": "inference.veo.fast.txt2vid.v2",
  "config": {
    "prompt": "A wooden water mill in a forest stream, water splashing on the wheel, leaves rustling, peaceful afternoon, cinematic HDR",
    "negative_prompt": "low quality, blurry, watermark",
    "resolution": "720p",
    "aspect_ratio": "16:9",
    "duration_seconds": 4,
    "generate_audio": true
  }
}
EOF
)

curl -sSf \
  -H "Authorization: Bearer $PRODIA_TOKEN" \
  -H 'Accept: video/mp4' \
  --json "$job" \
  --output watermill.mp4 \
  --retry 3 \
  https://inference.prodia.com/v2/job
```

```bash
bash main.sh
```

### macOS

```bash
open watermill.mp4
```

### Linux

```bash
xdg-open watermill.mp4
```

### Windows

```bash
start watermill.mp4
```

The output is a 4-second 720p MP4 with both video (H.264) and audio (AAC) tracks —
play it with sound on to hear the splashing water and rustling leaves the model
generated alongside the imagery.

![First frame of a Veo-generated video with audio — a wooden water mill in an autumn forest stream](https://docs.prodia.com/llms/assets/5bc1a674b73a-veo-with-audio-poster.jpg)

> tip:
>
> Audio quality follows the prompt. Mention specific sounds you want
> ("water splashing", "birds chirping", "footsteps on gravel") rather than relying
> on the model to infer them. For dialogue, put the spoken line in quotes.