# Wan 2.2 T2V Fast - Core.Today AI API

> A very fast and cheap PrunaAI-optimized version of Alibaba's Wan 2.2 A14B text-to-video model. Generates short clips at 480p or 720p directly from a text prompt, with 30 FPS frame interpolation enabled by default — the text-to-video sibling of Wan 2.2 I2V Fast.

- **Provider**: Alibaba
- **Model ID**: wan-video/wan-2.2-t2v-fast
- **Category**: Video Generation
- **Credits**: 240 per output video (flat, not per-second) — 480p 120, 720p 240 (billed at the 720p rate if resolution is omitted)
- **Speed**: Fast
- **Quality**: Standard

## Features
- PrunaAI-optimized inference — fast and low-cost compared to the base Wan 2.2 model
- Pure text-to-video: no reference image required, just a prompt
- 30 FPS frame interpolation (ffmpeg-based) enabled by default for smoother motion
- 480p or 720p output with flat per-video pricing (not per-second)
- 81-121 frame range and adjustable FPS for fine control over clip length

## Use Cases
- Quick text-to-video social media clips without any source footage
- Low-cost concept previews before committing to a higher-tier video model
- Batch-generating many prompt variations cheaply at 480p
- Vertical short-form content via the 9:16 aspect ratio
- Fast storyboard and animatic drafts from script lines

## API Endpoint

Base URL: https://api.core.today/v1
Create Prediction: POST /predictions
Get Status: GET /predictions/{job_id}

## Authentication
Header: X-API-Key: YOUR_API_KEY

## Input Parameters

### Required
- **prompt**: string - Prompt for video generation

### Optional
- **optimize_prompt**: boolean (default: false) - Translate prompt to Chinese before generation
- **num_frames**: integer (default: 81) - Number of video frames, 81-121 (81 gives the best results)
- **aspect_ratio**: string (default: 16:9) - Aspect ratio of the video
  Options: 16:9, 9:16
- **resolution**: string (default: 480p) - Resolution of video: 16:9 is 832x480px, 9:16 is 480x832px
  Options: 480p, 720p
- **frames_per_second**: integer (default: 16) - Frames per second, 5-30. Pricing is based on duration at 16 fps
- **interpolate_output**: boolean (default: true) - Interpolate the generated video to 30 FPS using ffmpeg (on by default)
- **go_fast**: boolean (default: true) - Go fast (faster inference, no price change)
- **sample_shift**: number (default: 12) - Sample shift factor (1-20)
- **seed**: integer - Random seed. Leave blank for random
- **disable_safety_checker**: boolean (default: false) - Disable safety checker for generated video
- **lora_weights_transformer**: string - Load LoRA weights for transformer. Supports arbitrary .safetensors URLs from the Internet (for example, 'https://huggingface.co/Viktor1717/scandinavian-interior-style1/resolve/main/my_first_flux_lora_v1.safetensors')
- **lora_scale_transformer**: number (default: 1) - Determines how strongly the transformer LoRA should be applied.
- **lora_weights_transformer_2**: string - Load LoRA weights for transformer_2. Supports arbitrary .safetensors URLs from the Internet. Can be different from transformer LoRA.
- **lora_scale_transformer_2**: number (default: 1) - Determines how strongly the transformer_2 LoRA should be applied.

## Examples

### 480p Text-to-Video (Half Price)
Fast, low-cost generation at the default 480p resolution — 75 credits
```json
{
  "model": "wan-video/wan-2.2-t2v-fast",
  "input": {
    "prompt": "A sports car is driving very fast along a beach at sunset",
    "go_fast": true,
    "num_frames": 81,
    "resolution": "480p",
    "aspect_ratio": "16:9",
    "sample_shift": 12,
    "frames_per_second": 16
  }
}
```

### 720p Vertical Clip
Higher-resolution 9:16 output for short-form social content
```json
{
  "model": "wan-video/wan-2.2-t2v-fast",
  "input": {
    "prompt": "A barista pours latte art in a cozy cafe, steam rising, warm morning light through the window",
    "resolution": "720p",
    "aspect_ratio": "9:16",
    "num_frames": 81,
    "frames_per_second": 16
  }
}
```

## Response Format
```json
{
  "job_id": "abc123",
  "status": "pending | processing | completed | failed",
  "result": "URL or data (when completed)"
}
```

## Usage Flow
1. POST /predictions with model and input -> receive job_id
2. GET /predictions/{job_id} -> poll until status is completed or failed
3. Result contains output URL(s)

## Tips
- 480p costs half of 720p (75 vs 150 credits) — prototype prompts at 480p and only re-render keepers at 720p
- Always set resolution explicitly: if it is omitted, billing conservatively falls back to the 720p rate even though the schema default is 480p
- Pricing is per output video, not per second — clip length is controlled by num_frames and frames_per_second without extra cost
- 81 frames gives the best quality according to the model card; going up to 121 rarely helps
- For animating an existing image instead of pure text, use the sibling model wan-video/wan-2.2-i2v-fast

## Documentation
https://replicate.com/wan-video/wan-2.2-t2v-fast