# GateAI Core API Documentation

> Unified AI API for image, video and 3D generation.

## General

**Base URL:** `https://core.claude.gg`

## Authentication

All requests (except `/docs.txt` and `/v1/models`) require an API key:

```
Authorization: Bearer sk-your-api-key-here
```

## Daily Budget

Each API key gets **$100/day** credit. Budget resets daily at **03:00 Turkey time** (midnight UTC).

**Example costs:**
- GPT Image 2 (1024×1024, LOW): ~$0.003
- GPT Image 2 (1024×1024, HIGH): ~$0.11
- Ideogram 3.0 (TURBO): ~$0.01
- Seedance 2.0 (720p, 4s): ~$0.49
- Seedance 2.0 (720p, 15s): ~$1.85
- Seedance 2.0 (1080p, 12s): ~$3.33
- Hailuo (720p, 6s): ~$0.04

---

## Endpoints

### Generate Content

```
POST https://core.claude.gg/api/generate
Authorization: Bearer sk-your-api-key
Content-Type: application/json
```

**Request body:**
```json
{
  "model": "model-name",
  "type": "text-to-image",
  "params": {
    "prompt": "A serene mountain lake at sunrise",
    "width": 1024,
    "height": 1024
  }
}
```

**Image/video references:** You can pass references directly in `params` as URLs or base64 strings (no pre-upload needed):

```json
{
  "model": "seedance-2.0",
  "params": {
    "prompt": "The cat stretches and walks away",
    "image": "https://example.com/cat.jpg",
    "mode": "RESOLUTION_720"
  }
}
```

Supported inline reference fields: `image`, `input_image`, `first_frame`, `last_frame`, `start_image_url`, `end_image_url`, `reference_images` (array), `image_url`, `image_urls` (array), `video_url`, `reference_video_url`, `audio_url`, `reference_audio_url`. Both HTTPS URLs and `data:` base64 URIs are accepted. Alternatively, you can pre-upload via `POST /api/upload` and send file IDs in the `images`/`videos` arrays.

**Response:**
```json
{
  "task_id": "f578692c-6475-4d85-b4c2-32a97af506ff",
  "url": "/get/f578692c-6475-4d85-b4c2-32a97af506ff",
  "status": "PROCESSING",
  "cost": 0.05,
  "remaining": 99.95
}
```

If the requested parameters exceed your daily budget, the server automatically downgrades resolution or shortens duration to fit. The `adjustments` field in the response shows what was changed. Use `/api/estimate` to preview cost before generating.

## Request Limits

- **Prompt length:** max 1500 characters (`PROMPT_TOO_LONG`)
- **Quantity per request:** max 8 images (`QUANTITY_TOO_HIGH`)
- **Concurrent requests:** max 5 in-flight per key (`RATE_LIMITED`)
- **Rate limit:** max 5 requests per 30s sliding window (`RATE_LIMITED`)

### Check Status

```
GET /get/{task_id}
```

**Response:**
```json
{
  "status": "FINISHED",
  "result": ["https://cdn.example.com/output.jpg"],
  "model": "gpt_image_2_0",
  "cost": 0.11
}
```

| Status | Meaning |
|---|---|
| PROCESSING | Generation in progress |
| FINISHED | Done, result URLs available |
| FAILED | Generation failed |

### Estimate Cost

```
POST https://core.claude.gg/api/estimate
Authorization: Bearer sk-your-api-key
Content-Type: application/json
```

Same request body as `/api/generate`. Returns cost in dollars without generating. If over budget, suggests adjusted parameters.

**Response:**
```json
{
  "cost": 0.11,
  "daily_remaining": 100.00,
  "fits_in_budget": true
}
```

### Model Discovery

Three endpoints for exploring available models and their parameters:

#### List All Models
```
GET /v1/models
```
Returns every available model with type (`image` / `video` / `3d`), cost summary, and parameter names. No auth required.

**Response:**
```json
[
  {
    "id": "gpt-image-2",
    "type": "image",
    "cost": "~$0.003/image (LOW)",
    "params": ["prompt", "width", "height", "quality", "quantity"]
  },
  {
    "id": "seedance-2.0",
    "type": "video",
    "cost": "~$0.49/720p 4s",
    "params": ["prompt", "mode", "duration", "motion_has_audio"]
  }
]
```

#### Full Schema (All Models)
```
GET /v1/models/schema
```
Returns the complete JSON schema for every model — parameter types, defaults, enums, min/max, descriptions. Useful for building dynamic UIs or validating requests client-side.

#### Single Model Schema
```
GET /v1/models/:modelId
```
Returns a single model's full schema + an example request body ready to copy-paste.

**Example:** `GET /v1/models/seedance-2.0`

**Response:**
```json
{
  "id": "seedance-2.0",
  "type": "video",
  "cost": "~$0.49/720p 4s, ~$1.85/720p 15s",
  "params": {
    "prompt": { "type": "string", "required": true, "description": "..." },
    "mode": { "type": "string", "enum": ["RESOLUTION_720", "RESOLUTION_1080"], "default": "RESOLUTION_720" },
    "duration": { "type": "integer", "min": 4, "max": 15, "default": 4 }
  },
  "example": {
    "model": "seedance-2.0",
    "params": { "prompt": "A cat walking", "mode": "RESOLUTION_720", "duration": 5 }
  }
}
```

**Fuzzy matching:** If the exact model ID is not found, the API suggests close matches. For example, `GET /v1/models/seedance` returns:
```json
{ "error": "Model not found", "suggestions": ["seedance-2.0"] }
```

### Check Credits

```
GET /credits
Authorization: Bearer sk-your-api-key
```

Returns your daily budget status and pool health.

**Response:**
```json
{
  "your_budget": {
    "daily_limit": 100,
    "used": 5.22,
    "remaining": 94.78,
  },
  "pool": { "total_accounts": 12, "locked": 1 }
}
```

### Personal Dashboard

```
GET /api/me?key=sk-your-api-key
```

Returns a full overview of your account: budget, active generations, concurrency slots, and spending breakdown. No `Authorization` header needed — pass your key as `?key=` query parameter.

**Response:**
```json
{
  "budget": {
    "daily_limit": 100,
    "used": 5.22,
    "remaining": 94.78
  },
  "concurrency": {
    "used": 2,
    "max": 5,
    "available": 3,
    "active_slots": [
      { "taskId": "abc-123", "model": "gpt-image-2", "elapsed": 12 },
      { "taskId": "def-456", "model": "seedance-2.0", "elapsed": 45 }
    ]
  },
  "active_tasks": [
    { "task_id": "abc-123", "model": "gpt-image-2", "status": "PROCESSING", "cost": 0.003, "elapsed": "12s" }
  ],
  "recent_tasks": [
    { "task_id": "old-789", "model": "gpt-image-2", "status": "FINISHED", "cost": 0.003, "result_count": 4 }
  ],
  "today_spending": {
    "gpt-image-2": { "requests": 15, "cost": 0.05 },
    "seedance-2.0": { "requests": 2, "cost": 3.70 }
  }
}
```

| Field | Description |
|---|---|
| `budget` | Daily limit, used, and remaining balance |
| `concurrency` | Active slots out of max 5 concurrent requests per key |
| `active_tasks` | Currently PROCESSING or QUEUED generations |
| `recent_tasks` | Last 20 completed generations (FINISHED/FAILED) |
| `today_spending` | Per-model cost breakdown for today |

**Rate limit:** Each API key can have up to **5 concurrent** generation requests. The 6th request returns `429 RATE_LIMITED` until a slot frees up. Concurrency slots auto-release after **10 minutes** even if the generation is still running — this only frees the slot, the generation itself continues.

**Generation timeouts:**

| Threshold | What happens |
|---|---|
| **20 minutes** | If the generation has not completed, your balance is **automatically refunded**. The generation keeps running — if it finishes after 20 min, you keep the result for free. |
| **40 minutes** | Hard deadline. The task is force-terminated with status `FAILED` and error `Generation timed out (40 min hard limit)`. |

These timeouts cover edge cases like slow video renders or upstream delays. Most generations complete well within 5 minutes (images: ~10-30s, videos: ~1-5 min).

---

## Cost Calculation

All costs are shown in **US dollars ($)**. Examples:

| Model | Parameters | Cost |
|---|---|---|
| GPT Image 2 | 1024×1024, LOW | ~$0.003 |
| GPT Image 2 | 1024×1024, HIGH | ~$0.11 |
| Ideogram 3.0 | TURBO | ~$0.01 |
| Seedance 2.0 | 720p, 4s | ~$0.49 |
| Seedance 2.0 | 720p, 15s | ~$1.85 |
| Seedance 2.0 | 1080p, 12s | ~$3.33 |
| Seedance 2.0 | 1080p, 15s | ~$11.25 |
| Hailuo | 720p, 6s | ~$0.04 |

Each API key has a **$100/day** budget that resets at **03:00 Turkey time** (midnight UTC) every day. If a generation would exceed your remaining budget, parameters are automatically downgraded (lower resolution or shorter duration). Use `/api/estimate` to preview cost.

---

## Image Generation Models

### Auto Image

`model: auto-image`

Cost: ~$0.02 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| width | integer | no | 1024 | - | - |
| height | integer | no | 1024 | - | - |
| prompt | ? | YES | - | - | - |
| quantity | integer | no | 1 | max: 8 | - |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 1:1 | 1024×1024 |
| 2:3 | 832×1248 |
| 3:2 | 1248×832 |
| 3:4 | 864×1184 |
| Twitter / X (4:3) | 1184×864 |
| Instagram (4:5) | 896×1152 |
| 5:4 | 1152×896 |
| 16:9 | 1344×768 |
| TikTok (9:16) | 768×1344 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "auto-image",
  "type": "image",
  "params": {
    "prompt": "A beautiful scene",
    "width": 1024,
    "height": 1024
  }
}
```

---

### Flux Dev

`model: flux-dev`

Cost: ~$0.00 per 1024×1024 image

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| mode | FAST / ULTRA | no | FAST | - | - |
| seed | integer | no | - | min: 0, max: 2147483637 | - |
| loras | array | no | - | - | LoRA models to apply during generation |
| width | integer | no | 1024 | min: 16, max: 2496, step: 8 | - |
| height | integer | no | 1024 | min: 16, max: 2496, step: 8 | - |
| prompt | string | YES | - | minLen: 1, maxLen: 2000 | The prompt to use for the generation |
| quantity | integer | no | 1 | min: 1, max: 8 | - |
| guidance_scale | number | no | 3.5 | min: 1, max: 10 | How closely the model follows the prompt |
| image_guidances | array | no | - | - | Image guidance configurations for StyleTransfer and LoraDepthMidas |
| num_inference_steps | integer | no | 28 | min: 1, max: 40 | Number of denoising steps |

**Dimension Range:**
- Width: 480–2048 (step: 8)
- Height: 480–2048 (step: 8)
- Invalid values are auto-snapped to the nearest valid step.

**Example:**
```json
{
  "model": "flux-dev",
  "type": "image",
  "params": {
    "prompt": "A beautiful scene",
    "mode": "FAST",
    "width": 1024,
    "height": 1024,
    "guidance_scale": 3.5,
    "num_inference_steps": 28
  }
}
```

---

### Flux Kontext Max

`model: flux-kontext-max`

Cost: ~$0.04 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| seed | integer | no | - | min: 0, max: 4294967295 | - |
| width | integer | no | 1024 | min: 32, max: 2048 | - |
| height | integer | no | 1024 | min: 32, max: 2048 | - |
| prompt | ? | YES | - | - | - |
| quantity | integer | no | 4 | max: 8 | - |
| prompt_enhance | ? | no | - | - | - |

**Dimension Range:**
- Width: 32–2048 (step: 1)
- Height: 32–2048 (step: 1)
- Invalid values are auto-snapped to the nearest valid step.

**Example:**
```json
{
  "model": "flux-kontext-max",
  "type": "image",
  "params": {
    "prompt": "A beautiful scene",
    "width": 1024,
    "height": 1024
  }
}
```

---

### Flux Kontext Pro

`model: flux-kontext-pro`

Cost: ~$0.02 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| seed | integer | no | - | min: 0, max: 2147483637 | - |
| width | integer | no | 1024 | min: 32, max: 2048 | - |
| height | integer | no | 1024 | min: 32, max: 2048 | - |
| prompt | ? | YES | - | - | - |
| quantity | integer | no | 4 | max: 8 | - |
| prompt_enhance | ? | no | - | - | - |

**Dimension Range:**
- Width: 32–2048 (step: 1)
- Height: 32–2048 (step: 1)
- Invalid values are auto-snapped to the nearest valid step.

**Example:**
```json
{
  "model": "flux-kontext-pro",
  "type": "image",
  "params": {
    "prompt": "A beautiful scene",
    "width": 1024,
    "height": 1024
  }
}
```

---

### Flux Pro 2.0

`model: flux-pro-2.0`

Cost: ~$0.01 per 1024×1024 image

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| seed | integer | no | - | min: 0, max: 4294967295 | - |
| width | integer | no | 1440 | min: 64, max: 14142 | - |
| height | integer | no | 1440 | min: 64, max: 14142 | - |
| prompt | ? | YES | - | - | - |
| quantity | integer | no | 4 | max: 8 | - |
| prompt_enhance | ? | no | - | - | - |

**Dimension Range:**
- Width: 64–14142 (step: 1)
- Height: 64–14142 (step: 1)
- Invalid values are auto-snapped to the nearest valid step.

**Example:**
```json
{
  "model": "flux-pro-2.0",
  "type": "image",
  "params": {
    "prompt": "A beautiful scene",
    "width": 1440,
    "height": 1440
  }
}
```

---

### Flux Schnell (1)

`model: flux-schnell`

Cost: ~$0.00 per 1024×1024 image

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| seed | integer | no | - | min: 0, max: 2147483637 | - |
| width | integer | no | 1024 | min: 32, max: 2048, step: 8 | - |
| height | integer | no | 1024 | min: 32, max: 2048, step: 8 | - |
| prompt | ? | YES | - | - | - |
| quantity | integer | no | 4 | max: 8 | - |
| prompt_enhance | ? | no | - | - | - |

**Dimension Range:**
- Width: 32–2048 (step: 8)
- Height: 32–2048 (step: 8)
- Invalid values are auto-snapped to the nearest valid step.

**Example:**
```json
{
  "model": "flux-schnell",
  "type": "image",
  "params": {
    "prompt": "A beautiful scene",
    "width": 1024,
    "height": 1024
  }
}
```

---

### Gemini 2.5 Flash (Nano Banana) Image

`model: gemini-2.5-flash-image`

Cost: ~$0.02 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| seed | integer | no | - | min: 0, max: 2147483638 | - |
| width | 0 / 768 / 832 / 864 / 896 / 1024 / 1152 / 1184 / 1248 / 1344 / 1536 | no | 1024 | - | - |
| height | 0 / 672 / 768 / 832 / 864 / 896 / 1024 / 1152 / 1184 / 1248 / 1344 | no | 1024 | - | - |
| prompt | ? | YES | - | - | - |
| quantity | integer | no | 1 | max: 8 | - |
| prompt_enhance | ? | no | - | - | - |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 1:1 | 1024×1024 |
| 2:3 | 832×1248 |
| 16:9 | 1344×768 |
| Twitter (4:3) | 1184×864 |
| Mobile (9:16) | 768×1344 |
| Instagram (4:5) | 896×1152 |
| 3:2 | 1248×832 |
| 3:4 | 864×1184 |
| 5:4 | 1152×896 |
| Ultrawide (21:9) | 1536×672 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "gemini-2.5-flash-image",
  "type": "image",
  "params": {
    "prompt": "A beautiful scene",
    "width": 1024,
    "height": 1024
  }
}
```

---

### Gemini Image 2

`model: gemini-image-2`

Cost: ~$0.06 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| seed | integer | no | - | min: 0, max: 2147483638 | Seed for reproducible image generation |
| top_p | number | no | - | min: 0, max: 1 | Nucleus sampling parameter (lower=more focused, higher=more diverse) |
| width | 768 / 848 / 896 / 928 / 1024 / 1152 / 1200 / 1264 / 1376 / 1536 / 1584 / 1696 / 1792 / 1856 / 2048 / 2304 / 2400 / 2528 / 2752 / 3072 / 3168 / 3392 / 3584 / 3712 / 4096 / 4608 / 4800 / 5056 / 5504 / 6336 | no | - | - | Width of the generated image |
| height | 672 / 768 / 848 / 896 / 928 / 1024 / 1152 / 1200 / 1264 / 1344 / 1376 / 1536 / 1696 / 1792 / 1856 / 2048 / 2304 / 2400 / 2528 / 2688 / 2752 / 3072 / 3392 / 3584 / 3712 / 4096 / 4608 / 4800 / 5056 / 5504 | no | - | - | Height of the generated image |
| prompt | string | YES | - | - | Text prompt for image generation |
| quantity | integer | no | 1 | min: 1, max: 8 | Number of images to generate |
| image_urls | ? | no | - | - | - |
| temperature | number | no | - | min: 0, max: 2 | Controls randomness in generation (0.0=deterministic, 2.0=very creative) |
| output_format | jpeg / png | no | png | - | Output image format |
| response_modalities | array | no | IMAGE | - | What types of content to generate |
| use_auto_dimensions | boolean | no | - | - | Internal flag indicating dimensions should not be sent to provider |

**Supported Resolutions (aspect ratio × size tier):**

| Aspect Ratio | Small | Medium | Large |
|---|---|---|---|
| 1:1 | 1024×1024 | 2048×2048 | 4096×4096 |
| 2:3 | 848×1264 | 1696×2528 | 3392×5056 |
| 16:9 | 1376×768 | 2752×1536 | 5504×3072 |
| Twitter (4:3) | 1200×896 | 2400×1792 | 4800×3584 |
| Mobile (9:16) | 768×1376 | 1536×2752 | 3072×5504 |
| Instagram (4:5) | 928×1152 | 1856×2304 | 3712×4608 |
| 3:2 | 1264×848 | 2528×1696 | 5056×3392 |
| 3:4 | 896×1200 | 1792×2400 | 3584×4800 |
| 5:4 | 1152×928 | 2304×1856 | 4608×3712 |
| Ultrawide (21:9) | 1584×672 | 3168×1344 | 6336×2688 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "gemini-image-2",
  "type": "image",
  "params": {
    "prompt": "A beautiful scene",
    "output_format": "png",
    "response_modalities": [
      "IMAGE"
    ]
  }
}
```

---

### GPT Image 1

`model: gpt-image-1`

Cost: ~$0.01 per 1024×1024 image

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| width | integer | no | 1024 | min: 1024, max: 1536, step: 8 | - |
| height | integer | no | 1024 | min: 1024, max: 1536, step: 8 | - |
| prompt | ? | YES | - | - | - |
| quality | high / medium / low | no | medium | - | The quality of the generated images. |
| quantity | integer | no | 1 | min: 1, max: 8 | - |
| transparency | boolean | no | - | - | - |
| output_format | jpeg / png / webp | no | jpeg | - | The format in which the generated images are returned. |
| image_guidances | array | no | - | - | - |

**Dimension Range:**
- Width: 32–2048 (step: 1)
- Height: 32–2048 (step: 1)
- Invalid values are auto-snapped to the nearest valid step.

**Example:**
```json
{
  "model": "gpt-image-1",
  "type": "image",
  "params": {
    "prompt": "A beautiful scene",
    "width": 1024,
    "height": 1024,
    "quality": "medium",
    "output_format": "jpeg"
  }
}
```

---

### GPT Image 1.5

`model: gpt-image-1.5`

Cost: ~$0.01 per 1024×1024 image

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| width | integer | no | 1024 | min: 1024, max: 1536, step: 8 | - |
| height | integer | no | 1024 | min: 1024, max: 1536, step: 8 | - |
| prompt | ? | YES | - | - | - |
| quality | high / medium / low | no | medium | - | The quality of the generated images. |
| quantity | integer | no | 1 | min: 1, max: 8 | - |
| output_format | jpeg / png / webp | no | jpeg | - | The format in which the generated images are returned. |
| image_guidances | array | no | - | - | - |

**Dimension Range:**
- Width: 1024–1536 (step: 8)
- Height: 1024–1536 (step: 8)
- Invalid values are auto-snapped to the nearest valid step.

**Example:**
```json
{
  "model": "gpt-image-1.5",
  "type": "image",
  "params": {
    "prompt": "A beautiful scene",
    "width": 1024,
    "height": 1024,
    "quality": "medium",
    "output_format": "jpeg"
  }
}
```

---

### GPT Image 2

`model: gpt-image-2`

Cost: ~$0.00 per 1024×1024 image

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| width | integer | no | 1024 | min: 256, max: 3840, step: 16 | Image width in pixels. Must be a multiple of 16. Each edge must be between 256 and 3840 px inclusive. Total pixels (widt |
| height | integer | no | 1024 | min: 256, max: 3840, step: 16 | Image height in pixels. Must be a multiple of 16. Each edge must be between 256 and 3840 px inclusive. Total pixels (wid |
| prompt | ? | YES | - | - | - |
| quality | high / medium / low | no | low | - | The quality of the generated images. |
| quantity | integer | no | 1 | min: 1, max: 8 | - |
| output_format | jpeg / png | no | jpeg | - | The format in which the generated images are returned. |
| image_guidances | array | no | - | - | - |

**Supported Resolutions (aspect ratio × size tier):**

| Aspect Ratio | Small | Medium | Large |
|---|---|---|---|
| 1:1 | 1024×1024 | 2048×2048 | 2880×2880 |
| 2:3 | 848×1264 | 1376×2048 | 2336×3504 |
| 16:9 | 1376×768 | 2048×1136 | 3584×2016 |
| Twitter (4:3) | 1200×896 | 2048×1536 | 3264×2448 |
| Mobile (9:16) | 768×1376 | 1136×2048 | 2016×3584 |
| Instagram (4:5) | 928×1152 | 1648×2048 | 2560×3200 |
| 3:2 | 1264×848 | 2048×1376 | 3504×2336 |
| 3:4 | 896×1200 | 1536×2048 | 2448×3264 |
| 5:4 | 1152×928 | 2048×1648 | 3200×2560 |
| Ultrawide (21:9) | 1584×672 | 2048×864 | 3808×1632 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "gpt-image-2",
  "type": "image",
  "params": {
    "prompt": "A beautiful scene",
    "width": 1024,
    "height": 1024,
    "quality": "low",
    "output_format": "jpeg"
  }
}
```

---

### Ideogram V3

`model: ideogram-v3.0`

Cost: ~$0.01 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| mode | TURBO / BALANCED / QUALITY | no | BALANCED | - | - |
| seed | integer | no | - | min: 0, max: 2147483637 | - |
| style | ? | no | - | - | - |
| width | integer | no | 1024 | min: 512, max: 1536, step: 8 | - |
| height | integer | no | 1024 | min: 512, max: 1536, step: 8 | - |
| prompt | string | YES | - | - | - |
| quantity | integer | no | 1 | min: 1, max: 8 | - |
| image_urls | ? | no | - | - | - |
| prompt_enhance | boolean | no | false | - | - |
| negative_prompt | string | no |  | - | - |

**Dimension Range:**
- Width: 32–2048 (step: 1)
- Height: 32–2048 (step: 1)
- Invalid values are auto-snapped to the nearest valid step.

**Example:**
```json
{
  "model": "ideogram-v3.0",
  "type": "image",
  "params": {
    "prompt": "A beautiful scene",
    "mode": "BALANCED",
    "width": 1024,
    "height": 1024,
    "prompt_enhance": false,
    "negative_prompt": ""
  }
}
```

---

### 'Cinematic Kino' (Kino XL)

`model: kino-xl`

Cost: ~$0.00 per 1024×1024 image

**Example:**
```json
{
  "model": "kino-xl",
  "type": "image",
  "params": {}
}
```

---

### Cortex 1.0

`model: cortex_1.0`

Cost: ~$0.01 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| quantity | number | no | 1 | min: 1, max: 1 | - |
| motion_strength | ? | no | - | - | - |

**Example:**
```json
{
  "model": "cortex_1.0",
  "type": "image",
  "params": {}
}
```

---

### Nano Banana 2

`model: nano-banana-2`

Cost: ~$0.03 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| seed | integer | no | - | min: 0, max: 2147483638 | Seed for reproducible image generation |
| top_p | number | no | - | min: 0, max: 1 | Nucleus sampling parameter (lower=more focused, higher=more diverse) |
| width | 768 / 848 / 896 / 928 / 1024 / 1152 / 1200 / 1264 / 1376 / 1536 / 1584 / 1696 / 1792 / 1856 / 2048 / 2304 / 2400 / 2528 / 2752 / 3072 / 3168 / 3392 / 3584 / 3712 / 4096 / 4608 / 4800 / 5056 / 5504 / 6336 | no | - | - | Width of the generated image |
| height | 672 / 768 / 848 / 896 / 928 / 1024 / 1152 / 1200 / 1264 / 1344 / 1376 / 1536 / 1696 / 1792 / 1856 / 2048 / 2304 / 2400 / 2528 / 2688 / 2752 / 3072 / 3392 / 3584 / 3712 / 4096 / 4608 / 4800 / 5056 / 5504 | no | - | - | Height of the generated image |
| prompt | string | YES | - | - | Text prompt for image generation |
| quantity | integer | no | 1 | min: 1, max: 8 | Number of images to generate |
| image_urls | ? | no | - | - | - |
| temperature | number | no | - | min: 0, max: 2 | Controls randomness in generation (0.0=deterministic, 2.0=very creative) |
| output_format | jpeg / png | no | png | - | Output image format |
| response_modalities | array | no | IMAGE | - | What types of content to generate |
| use_auto_dimensions | boolean | no | - | - | Internal flag indicating dimensions should not be sent to provider |

**Supported Resolutions (aspect ratio × size tier):**

| Aspect Ratio | Small | Medium | Large |
|---|---|---|---|
| 1:1 | 1024×1024 | 2048×2048 | 4096×4096 |
| 2:3 | 848×1264 | 1696×2528 | 3392×5056 |
| 16:9 | 1376×768 | 2752×1536 | 5504×3072 |
| Twitter (4:3) | 1200×896 | 2400×1792 | 4800×3584 |
| Mobile (9:16) | 768×1376 | 1536×2752 | 3072×5504 |
| Instagram (4:5) | 928×1152 | 1856×2304 | 3712×4608 |
| 3:2 | 1264×848 | 2528×1696 | 5056×3392 |
| 3:4 | 896×1200 | 1792×2400 | 3584×4800 |
| 5:4 | 1152×928 | 2304×1856 | 4608×3712 |
| Ultrawide (21:9) | 1584×672 | 3168×1344 | 6336×2688 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "nano-banana-2",
  "type": "image",
  "params": {
    "prompt": "A beautiful scene",
    "output_format": "png",
    "response_modalities": [
      "IMAGE"
    ]
  }
}
```

---

### Recraft V4

`model: recraft-v4`

Cost: ~$0.02 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| width | 768 / 832 / 896 / 1024 / 1152 / 1216 / 1280 / 1344 / 1536 | no | 1024 | - | - |
| height | 768 / 832 / 896 / 1024 / 1152 / 1216 / 1280 / 1344 / 1536 | no | 1024 | - | - |
| prompt | ? | YES | - | - | - |
| quantity | integer | no | 4 | max: 8 | - |
| prompt_enhance | ? | no | - | - | - |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 1:1 | 1024×1024 |
| 2:3 | 832×1280 |
| 16:9 | 1344×768 |
| Twitter (4:3) | 1216×896 |
| Mobile (9:16) | 768×1344 |
| Instagram (4:5) | 896×1152 |
| 1:2 | 768×1536 |
| 2:1 | 1536×768 |
| 3:2 | 1280×832 |
| 3:4 | 896×1216 |
| 5:4 | 1152×896 |
| 6:10 | 832×1344 |
| 10:14 | 896×1280 |
| 14:10 | 1280×896 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "recraft-v4",
  "type": "image",
  "params": {
    "prompt": "A beautiful scene",
    "width": 1024,
    "height": 1024
  }
}
```

---

### Recraft V4 Pro

`model: recraft-v4-pro`

Cost: ~$0.10 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| width | 1536 / 1664 / 1792 / 2048 / 2304 / 2432 / 2560 / 2688 / 3072 | no | 2048 | - | - |
| height | 1536 / 1664 / 1792 / 2048 / 2304 / 2432 / 2560 / 2688 / 3072 | no | 2048 | - | - |
| prompt | ? | YES | - | - | - |
| quantity | integer | no | 4 | max: 8 | - |
| prompt_enhance | ? | no | - | - | - |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 1:1 | 2048×2048 |
| 2:3 | 1664×2560 |
| 16:9 | 2688×1536 |
| Twitter (4:3) | 2432×1792 |
| Mobile (9:16) | 1536×2688 |
| Instagram (4:5) | 1792×2304 |
| 1:2 | 1536×3072 |
| 2:1 | 3072×1536 |
| 3:2 | 2560×1664 |
| 3:4 | 1792×2432 |
| 5:4 | 2304×1792 |
| 6:10 | 1664×2688 |
| 10:14 | 1792×2560 |
| 14:10 | 2560×1792 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "recraft-v4-pro",
  "type": "image",
  "params": {
    "prompt": "A beautiful scene",
    "width": 2048,
    "height": 2048
  }
}
```

---

### Seedream 4.0 Image

`model: seedream-4.0`

Cost: ~$0.01 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| seed | integer | no | - | min: 0, max: 2147483638 | - |
| width | integer | no | 1024 | min: 32, max: 14142 | - |
| height | integer | no | 1024 | min: 32, max: 14142 | - |
| prompt | ? | YES | - | - | - |
| quantity | integer | no | 1 | max: 8 | - |
| prompt_enhance | ? | no | - | - | - |

**Dimension Range:**
- Width: 32–14142 (step: 1)
- Height: 32–14142 (step: 1)
- Invalid values are auto-snapped to the nearest valid step.

**Example:**
```json
{
  "model": "seedream-4.0",
  "type": "image",
  "params": {
    "prompt": "A beautiful scene",
    "width": 1024,
    "height": 1024
  }
}
```

---

### Seedream 4.5 Image

`model: seedream-4.5`

Cost: ~$0.01 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| seed | integer | no | - | min: 0, max: 2147483638 | - |
| width | integer | no | 2048 | min: 32, max: 14142 | - |
| height | integer | no | 2048 | min: 32, max: 14142 | - |
| prompt | ? | YES | - | - | - |
| quantity | integer | no | 1 | max: 8 | - |
| prompt_enhance | ? | no | - | - | - |

**Dimension Range:**
- Width: 32–14142 (step: 1)
- Height: 32–14142 (step: 1)
- Invalid values are auto-snapped to the nearest valid step.

**Example:**
```json
{
  "model": "seedream-4.5",
  "type": "image",
  "params": {
    "prompt": "A beautiful scene",
    "width": 2048,
    "height": 2048
  }
}
```

---

## Video Generation Models

### Auto Video

`model: auto-video`

Cost: ~$0.13 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| width | integer | no | 1280 | - | - |
| height | integer | no | 720 | - | - |
| prompt | ? | YES | - | - | - |
| duration | 6 / 8 | no | 8 | - | - |
| quantity | integer | no | 1 | max: 1 | - |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 16:9 | 1280×720 |
| 9:16 | 720×1280 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "auto-video",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "width": 1280,
    "height": 720,
    "duration": 8
  }
}
```

---

### Hailuo-2.3 Video Gen

`model: hailuo-2_3`

Cost: ~$0.04 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| prompt | ? | YES | - | - | Text description for video generation |
| quantity | integer | no | 1 | min: 1, max: 1 | Number of videos to generate |
| resolution | 768p / 1080p | no | 1080p | - | Video resolution |
| start_frame | ? | no | - | - | First frame image for image-to-video generation |
| durationSeconds | 6 / 10 | no | 6 | - | Video duration in seconds. 1080P supports only 6 seconds |
| prompt_optimizer | boolean | no | true | - | Auto-optimize prompt for better video quality |
| fast_pretreatment | boolean | no | false | - | Reduce processing time during prompt optimization |

**Valid Widths:** 768, 1376, 1080, 1920
**Valid Heights:** 768, 1376, 1080, 1920

**Example:**
```json
{
  "model": "hailuo-2_3",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "resolution": "1080p",
    "durationSeconds": 6,
    "prompt_optimizer": true,
    "fast_pretreatment": false
  }
}
```

---

### Hailuo-2.3-fast Video Gen

`model: hailuo-2_3-fast`

Cost: ~$0.03 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| prompt | ? | YES | - | - | Text description for video generation |
| quantity | integer | no | 1 | min: 1, max: 1 | Number of videos to generate |
| resolution | 768p / 1080p | no | 1080p | - | Video resolution |
| start_frame | ? | YES | - | - | First frame image for image-to-video generation (required for fast variant) |
| durationSeconds | 6 / 10 | no | 6 | - | Video duration in seconds. 1080P supports only 6 seconds |
| prompt_optimizer | boolean | no | true | - | Auto-optimize prompt for better video quality |
| fast_pretreatment | boolean | no | false | - | Reduce processing time during prompt optimization |

**Valid Widths:** 768, 1376, 1080, 1920
**Valid Heights:** 768, 1376, 1080, 1920

**Example:**
```json
{
  "model": "hailuo-2_3-fast",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "resolution": "1080p",
    "durationSeconds": 6,
    "prompt_optimizer": true,
    "fast_pretreatment": false
  }
}
```

---

### Happy Horse Video Gen

`model: happy-horse`

Cost: ~$0.06 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| seed | integer | no | - | min: 0, max: 2147483647 | Random seed for reproducibility |
| width | integer | no | 1920 | - | Video width in pixels |
| height | integer | no | 1080 | - | Video height in pixels |
| prompt | ? | no | - | maxLen: 2500 | Text prompt describing the desired video |
| duration | 3 / 4 / 5 / 6 / 7 / 8 / 9 / 10 / 11 / 12 / 13 / 14 / 15 | no | 5 | - | Output video duration in seconds. Not applicable for video-edit (inherits from source video). |
| quantity | integer | no | 1 | min: 1, max: 1 | Number of videos to generate |
| start_frame | ? | no | - | - | Image to use as first frame for image-to-video generation |
| audio_setting | auto / origin | no | auto | - | Audio handling for video-edit: auto (model decides) or origin (preserve input audio). |
| image_guidance | array | no | - | - | Reference images for reference-to-video (1-9, referenced as character1-character9) or video-edit (0-5, referenced as @Im |
| video_guidance | array | no | - | - | Source video for video-edit mode. Maximum 1 item. |
| enable_safety_checker | boolean | no | false | - | Enable content moderation for input and output |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 1:1 | 1080×1080 |
| 3:4 | 1080×1440 |
| 4:3 | 1440×1080 |
| 16:9 | 1920×1080 |
| 9:16 | 1080×1920 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "happy-horse",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "width": 1920,
    "height": 1080,
    "duration": 5,
    "audio_setting": "auto",
    "enable_safety_checker": false
  }
}
```

---

### Kling 2.1 Pro

`model: kling-2.1-pro`

Cost: ~$0.04 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| width | integer | no | 1920 | - | - |
| height | integer | no | 1080 | - | - |
| prompt | ? | YES | - | - | - |
| duration | 5 / 10 | no | 5 | - | - |
| quantity | number | no | 1 | max: 1 | - |
| prompt_enhance | ? | no | - | - | - |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 1:1 | 1440×1440 |
| 16:9 | 1920×1080 |
| 9:16 | 1080×1920 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "kling-2.1-pro",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "width": 1920,
    "height": 1080,
    "duration": 5
  }
}
```

---

### Kling 2.5 Turbo

`model: kling-2.5`

Cost: ~$0.03 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| mode | RESOLUTION_720 / RESOLUTION_1080 | no | RESOLUTION_1080 | - | - |
| width | integer | no | 1920 | - | - |
| height | integer | no | 1080 | - | - |
| prompt | ? | YES | - | - | - |
| duration | 5 / 10 | no | 5 | - | - |
| quantity | number | no | 1 | max: 1 | - |
| prompt_enhance | ? | no | - | - | - |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 1:1 | 1440×1440 |
| 16:9 | 1920×1080 |
| 9:16 | 1080×1920 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "kling-2.5",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "mode": "RESOLUTION_1080",
    "width": 1920,
    "height": 1080,
    "duration": 5
  }
}
```

---

### Kling 2.5 Turbo Standard Video Gen

`model: kling-2.5-turbo-standard`

Cost: ~$0.02 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| width | 1280 / 720 / 960 | no | 1280 | - | Video width in pixels |
| height | 720 / 1280 / 960 | no | 720 | - | Video height in pixels |
| prompt | ? | YES | - | maxLen: 2500 | Text description for video generation |
| quantity | integer | no | 1 | min: 1, max: 1 | Number of videos to generate |
| cfg_scale | number | no | 0.5 | min: 0, max: 1 | Guidance scale for generation strength |
| start_frame | ? | YES | - | - | Start frame image for video generation |
| durationSeconds | 5 / 10 | no | 5 | - | Video duration in seconds |
| negative_prompt | string | no | blur, distort, and low quality | maxLen: 1000 | Elements to exclude from generation |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 1:1 | 960×960 |
| 16:9 | 1280×720 |
| 9:16 | 720×1280 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "kling-2.5-turbo-standard",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "width": 1280,
    "height": 720,
    "cfg_scale": 0.5,
    "durationSeconds": 5,
    "negative_prompt": "blur, distort, and low quality"
  }
}
```

---

### Kling 2.6

`model: kling-2.6`

Cost: ~$0.06 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| audio | ? | no | - | - | - |
| width | 1920 / 1080 / 1440 / 0 | no | - | - | - |
| height | 1080 / 1920 / 1440 / 0 | no | - | - | - |
| prompt | ? | YES | - | - | - |
| duration | 5 / 10 | no | 5 | - | - |
| quantity | number | no | 1 | max: 1 | - |
| start_frame | ? | no | - | - | - |
| motion_has_audio | ? | no | true | - | - |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 1:1 | 1440×1440 |
| 16:9 | 1920×1080 |
| 9:16 | 1080×1920 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "kling-2.6",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "duration": 5,
    "motion_has_audio": true
  }
}
```

---

### Kling 3.0 Video Gen

`model: kling-3.0`

Cost: ~$0.03 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| mode | RESOLUTION_720 / RESOLUTION_1080 / RESOLUTION_2160 | no | RESOLUTION_1080 | - | Video resolution mode. RESOLUTION_1080 uses pro endpoints (1080p), RESOLUTION_720 uses standard endpoints (720p), RESOLU |
| width | 1920 / 1080 / 1440 / 1280 / 720 / 960 / 0 / 3840 / 2160 / 2880 | no | 1920 | - | Video width in pixels |
| height | 1080 / 1920 / 1440 / 720 / 1280 / 960 / 0 / 2160 / 3840 / 2880 | no | 1080 | - | Video height in pixels |
| prompt | ? | YES | - | - | Text description for video generation. Can include dialogue, narration, sound effects, and ambient audio descriptions. |
| elements | array | no | - | - | Elements (characters/objects) to include. Reference in prompt as @Element1, @Element2. Supports element consistency for  |
| quantity | integer | no | 1 | min: 1, max: 1 | Number of videos to generate |
| end_frame | ? | no | - | - | End frame image for frame interpolation. Cannot be used when more than 2 images are used (start_frame + reference images |
| voice_ids | array | no | - | - | Voice IDs for voice control. Reference in prompt using <<<voice_1>>>, <<<voice_2>>> |
| image_urls | array | no | - | - | Reference images for style/appearance. Reference in prompt as @Image1, @Image2, etc. Maximum 7 total (elements + referen |
| start_frame | ? | no | - | - | Start frame image for image-to-video mode. Accepted formats: jpg, jpeg, png, webp, gif, avif. Max size: 10MB. Min dimens |
| generate_audio | boolean | no | true | - | Whether to generate native audio for the video. Supports multiple audio types: speech, dialogue, singing/rap, ambient, s |
| durationSeconds | 3 / 4 / 5 / 6 / 7 / 8 / 9 / 10 / 11 / 12 / 13 / 14 / 15 | no | 5 | - | Video duration in seconds (3-15s) |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 1:1 | 1440×1440 |
| 16:9 | 1920×1080 |
| 9:16 | 1080×1920 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "kling-3.0",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "mode": "RESOLUTION_1080",
    "width": 1920,
    "height": 1080,
    "generate_audio": true,
    "durationSeconds": 5
  }
}
```

---

### Kling Video O 1 Gen

`model: kling-video-o-1`

Cost: ~$0.04 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| width | integer | no | - | min: 0, max: 2160 | Video width in pixels (0 for auto) |
| height | integer | no | - | min: 0, max: 2160 | Video height in pixels (0 for auto) |
| prompt | ? | YES | - | - | Text description for video generation |
| quantity | integer | no | 1 | min: 1, max: 1 | Number of videos to generate |
| cfg_scale | number | no | 0.5 | min: 0, max: 1 | Guidance scale for generation strength |
| end_frame | ? | no | - | - | - |
| keep_audio | boolean | no | false | - | Whether to keep the original audio from the video. Defaults to false (matches Fal API default). Can be set to true to pr |
| start_frame | ? | no | - | - | - |
| image_guidance | array | no | - | - | Additional reference images for guidance. Up to 6 guidance images allowed (7 max allowed without videos, but start frame |
| video_guidance | array | no | - | - | Video guidance for video-to-video editing. Maximum 1 video allowed. When using BASE type, routes to video-to-video edit  |
| durationSeconds | 3 / 4 / 5 / 6 / 7 / 8 / 9 / 10 | no | 5 | - | Video duration in seconds |

**Dimension Range:**
- Width: 0–2160 (step: 1)
- Height: 0–2160 (step: 1)
- Invalid values are auto-snapped to the nearest valid step.

**Example:**
```json
{
  "model": "kling-video-o-1",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "cfg_scale": 0.5,
    "keep_audio": false,
    "durationSeconds": 5
  }
}
```

---

### Kling Video O 3 Gen

`model: kling-video-o-3`

Cost: ~$0.07 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| mode | RESOLUTION_720 / RESOLUTION_1080 / RESOLUTION_2160 | no | RESOLUTION_1080 | - | Video resolution mode. RESOLUTION_1080 uses pro endpoints (1080p), RESOLUTION_720 uses standard endpoints (720p), RESOLU |
| width | integer | no | - | min: 0, max: 3840 | Video width in pixels (0 for auto) |
| height | integer | no | - | min: 0, max: 3840 | Video height in pixels (0 for auto) |
| prompt | ? | YES | - | - | Text description for video generation |
| quantity | integer | no | 1 | min: 1, max: 1 | Number of videos to generate |
| end_frame | ? | no | - | - | End frame image for image-to-video generation |
| keep_audio | boolean | no | false | - | Whether to preserve audio from the reference video (video-to-video only) |
| start_frame | ? | no | - | - | Start frame image for image-to-video generation |
| generate_audio | boolean | no | false | - | Whether to generate native audio for the video |
| image_guidance | array | no | - | - | Reference images for style/appearance. Reference in prompt as @Image1, @Image2, etc. |
| video_guidance | array | no | - | - | Video guidance for video-to-video transformations. BASE type routes to edit endpoint, FEATURE type routes to reference e |
| durationSeconds | 3 / 4 / 5 / 6 / 7 / 8 / 9 / 10 / 11 / 12 / 13 / 14 / 15 | no | 5 | - | Video duration in seconds (3-15s) |

**Dimension Range:**
- Width: 0–3840 (step: 1)
- Height: 0–3840 (step: 1)
- Invalid values are auto-snapped to the nearest valid step.

**Example:**
```json
{
  "model": "kling-video-o-3",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "mode": "RESOLUTION_1080",
    "keep_audio": false,
    "generate_audio": false,
    "durationSeconds": 5
  }
}
```

---

### LTXV 2 Fast Video Gen

`model: ltxv-2.0-fast`

Cost: ~$0.02 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| fps | 25 / 50 | no | 50 | - | Frames per second |
| width | integer | no | - | - | Video width in pixels |
| height | integer | no | - | - | Video height in pixels |
| prompt | ? | YES | - | - | Text description for video generation |
| quantity | integer | no | 1 | min: 1, max: 1 | Number of videos to generate |
| resolution | 1080p / 1440p / 2160p | no | 1080p | - | Video resolution |
| start_frame | ? | no | - | - | Start frame image for image-to-video generation |
| aspect_ratio | 16:9 / 9:16 | no | 16:9 | - | Video aspect ratio |
| durationSeconds | 6 / 8 | no | 6 | - | Video duration in seconds |
| enable_prompt_expansion | boolean | no | true | - | Enable automatic prompt expansion |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 16:9 | 1920×1080 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "ltxv-2.0-fast",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "fps": 50,
    "resolution": "1080p",
    "aspect_ratio": "16:9",
    "durationSeconds": 6,
    "enable_prompt_expansion": true
  }
}
```

---

### LTXV 2 Pro Video Gen

`model: ltxv-2.0-pro`

Cost: ~$0.02 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| fps | 24 / 30 / 50 | no | 50 | - | Frames per second |
| width | integer | no | - | - | Video width in pixels |
| height | integer | no | - | - | Video height in pixels |
| prompt | ? | YES | - | - | Text description for video generation |
| quantity | integer | no | 1 | min: 1, max: 1 | Number of videos to generate |
| resolution | 1080p / 1440p / 2160p | no | 1080p | - | Video resolution |
| start_frame | ? | no | - | - | Start frame image for image-to-video generation |
| aspect_ratio | 16:9 / 9:16 | no | 16:9 | - | Video aspect ratio |
| durationSeconds | 6 / 8 | no | 6 | - | Video duration in seconds |
| enable_prompt_expansion | boolean | no | true | - | Enable automatic prompt expansion |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 16:9 | 1920×1080 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "ltxv-2.0-pro",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "fps": 50,
    "resolution": "1080p",
    "aspect_ratio": "16:9",
    "durationSeconds": 6,
    "enable_prompt_expansion": true
  }
}
```

---

### LTX 2.0 Ultra

`model: ltxv-2.0-ultra`

Cost: ~$0.02 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| mode | RESOLUTION_1080 / RESOLUTION_1440 / RESOLUTION_2160 | no | RESOLUTION_1080 | - | - |
| width | integer | YES | 1920 | - | - |
| height | integer | YES | 1080 | - | - |
| prompt | ? | YES | - | - | - |
| duration | 6 / 8 | no | 8 | - | - |
| quantity | integer | YES | 1 | max: 1 | - |
| prompt_enhance | ? | no | - | - | - |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 16:9 | 1920×1080 |
| 9:16 | 1080×1920 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "ltxv-2.0-ultra",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "mode": "RESOLUTION_1080",
    "width": 1920,
    "height": 1080,
    "duration": 8
  }
}
```

---

### LTXV 2.3 Fast Video Gen

`model: ltxv-2.3-fast`

Cost: ~$0.02 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| fps | 24 / 25 / 48 / 50 | no | 25 | - | Frames per second |
| width | integer | no | - | - | Video width in pixels |
| height | integer | no | - | - | Video height in pixels |
| prompt | ? | YES | - | - | Text description for video generation |
| quantity | integer | no | 1 | min: 1, max: 1 | Number of videos to generate |
| end_frame | ? | no | - | - | End frame image for transition video generation. Requires start_frame. |
| resolution | 1080p / 1440p / 2160p | no | 1080p | - | Video resolution |
| start_frame | ? | no | - | - | Start frame image for image-to-video generation |
| aspect_ratio | 16:9 / 9:16 | no | - | - | Video aspect ratio. Defaults to 16:9 for T2V, or auto-detected from start_frame for I2V. |
| generate_audio | boolean | no | true | - | Whether to generate audio for the video |
| durationSeconds | 6 / 8 / 10 / 12 / 14 / 16 / 18 / 20 | no | 6 | - | Video duration in seconds. Durations over 10s require 25fps and 1080p resolution. |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 16:9 | 1920×1080 |
| 9:16 | 1080×1920 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "ltxv-2.3-fast",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "fps": 25,
    "resolution": "1080p",
    "generate_audio": true,
    "durationSeconds": 6
  }
}
```

---

### LTXV 2.3 Pro Video Gen

`model: ltxv-2.3-pro`

Cost: ~$0.02 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| fps | 24 / 25 / 48 / 50 | no | 25 | - | Frames per second |
| width | integer | no | - | - | Video width in pixels |
| height | integer | no | - | - | Video height in pixels |
| prompt | ? | YES | - | - | Text description for video generation |
| quantity | integer | no | 1 | min: 1, max: 1 | Number of videos to generate |
| end_frame | ? | no | - | - | End frame image for transition video generation. Requires start_frame. |
| resolution | 1080p / 1440p / 2160p | no | 1080p | - | Video resolution |
| start_frame | ? | no | - | - | Start frame image for image-to-video generation |
| aspect_ratio | 16:9 / 9:16 | no | - | - | Video aspect ratio. Defaults to 16:9 for T2V, or auto-detected from start_frame for I2V. |
| generate_audio | boolean | no | true | - | Whether to generate audio for the video |
| durationSeconds | 6 / 8 / 10 | no | 6 | - | Video duration in seconds |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 16:9 | 1920×1080 |
| 9:16 | 1080×1920 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "ltxv-2.3-pro",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "fps": 25,
    "resolution": "1080p",
    "generate_audio": true,
    "durationSeconds": 6
  }
}
```

---

### Cortex 1.5

`model: lucid-origin`

Cost: ~$0.00 per 1024×1024 image

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| mode | FAST / ULTRA | no | FAST | - | - |
| seed | integer | no | - | min: 0, max: 2147483637 | - |
| width | integer | no | 1024 | min: 16, max: 3000, step: 8 | - |
| height | integer | no | 1024 | min: 16, max: 3000, step: 8 | - |
| prompt | string | YES | - | minLen: 1, maxLen: 2000 | The prompt to use for the generation |
| quantity | integer | no | 1 | min: 1, max: 8 | - |
| guidance_scale | number | no | 0 | min: 0, max: 20 | How closely the model follows the prompt |
| image_guidances | array | no | - | - | Image guidance configurations for StyleTransfer and LoraDepthMidas |
| num_inference_steps | integer | no | 15 | min: 1, max: 150 | Number of denoising steps |

**Dimension Range:**
- Width: 16–3840 (step: 8)
- Height: 16–3616 (step: 8)
- Invalid values are auto-snapped to the nearest valid step.

**Example:**
```json
{
  "model": "lucid-origin",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "mode": "FAST",
    "width": 1024,
    "height": 1024,
    "guidance_scale": 0,
    "num_inference_steps": 15
  }
}
```

---

### Lucid Realism

`model: lucid-realism`

Cost: ~$0.00 per 1024×1024 image

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| mode | FAST / ULTRA | no | FAST | - | - |
| seed | integer | no | - | min: 0, max: 2147483647 | - |
| width | integer | no | 1024 | min: 16, max: 2496, step: 8 | - |
| height | integer | no | 1024 | min: 16, max: 2496, step: 8 | - |
| prompt | ? | YES | - | - | - |
| quantity | integer | no | 4 | max: 8 | - |
| prompt_enhance | ? | no | - | - | - |

**Dimension Range:**
- Width: 16–2496 (step: 8)
- Height: 16–2496 (step: 8)
- Invalid values are auto-snapped to the nearest valid step.

**Example:**
```json
{
  "model": "lucid-realism",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "mode": "FAST",
    "width": 1024,
    "height": 1024
  }
}
```

---

### Cortex 2.0

`model: cortex_2.0`

Cost: ~$0.08 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| mode | RESOLUTION_480 / RESOLUTION_720 | no | RESOLUTION_480 | - | - |
| seed | number | no | - | min: 0, max: 2147483637 | - |
| width | ? | no | - | - | - |
| height | ? | no | - | - | - |
| prompt | ? | YES | - | - | - |
| controls | ? | no | - | - | - |
| elements | ? | no | - | - | - |
| quantity | number | no | 1 | min: 1, max: 1 | - |
| prompt_enhance | ? | no | - | - | - |
| negative_prompt | ? | no | - | - | - |
| frame_interpolation | ? | no | - | - | - |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 2:3 | 512×768 |
| 4:5 | 576×720 |
| 16:9 | 832×480 |
| 9:16 | 480×832 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "cortex_2.0",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "mode": "RESOLUTION_480"
  }
}
```

---

### Cortex 2.0 Fast

`model: cortex_2.0-fast`

Cost: ~$0.03 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| mode | RESOLUTION_480 / RESOLUTION_720 | no | RESOLUTION_480 | - | - |
| seed | number | no | - | min: 0, max: 4294967295 | - |
| width | ? | no | - | - | - |
| height | ? | no | - | - | - |
| prompt | ? | YES | - | - | - |
| controls | ? | no | - | - | - |
| elements | ? | no | - | - | - |
| quantity | number | no | 1 | min: 1, max: 1 | - |
| start_frame | ? | no | - | - | - |
| prompt_enhance | ? | no | - | - | - |
| negative_prompt | ? | no | - | - | - |
| frame_interpolation | ? | no | - | - | - |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 2:3 | 512×768 |
| 4:5 | 576×720 |
| 16:9 | 832×480 |
| 9:16 | 480×832 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "cortex_2.0-fast",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "mode": "RESOLUTION_480"
  }
}
```

---

### Cortex v0.9

`model: cortex-v0.9`

Cost: ~$0.00 per 1024×1024 image (min $0.00)

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| mode | FAST / QUALITY / ULTRA | no | FAST | - | - |
| seed | integer | no | - | min: 0 | - |
| width | integer | no | 1024 | min: 32, max: 2048, step: 8 | - |
| height | integer | no | 1024 | min: 32, max: 2048, step: 8 | - |
| prompt | ? | YES | - | - | - |
| tiling | ? | no | - | - | - |
| contrast | ? | no | - | - | - |
| quantity | integer | no | 4 | max: 8 | - |
| prompt_enhance | ? | no | - | - | - |
| negative_prompt | ? | no | - | - | - |

**Dimension Range:**
- Width: 32–2048 (step: 8)
- Height: 32–2048 (step: 8)
- Invalid values are auto-snapped to the nearest valid step.

**Example:**
```json
{
  "model": "cortex-v0.9",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "mode": "FAST",
    "width": 1024,
    "height": 1024
  }
}
```

---

### Cortex v1.0

`model: cortex-v1.0`

Cost: ~$0.00 per 1024×1024 image (min $0.00)

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| mode | FAST / QUALITY / ULTRA | no | FAST | - | - |
| seed | integer | no | - | min: 0 | - |
| width | integer | no | 1024 | min: 32, max: 2048, step: 8 | - |
| height | integer | no | 1024 | min: 32, max: 2048, step: 8 | - |
| prompt | ? | YES | - | - | - |
| tiling | ? | no | - | - | - |
| contrast | ? | no | - | - | - |
| quantity | integer | no | 4 | max: 8 | - |
| prompt_enhance | ? | no | - | - | - |
| negative_prompt | ? | no | - | - | - |

**Dimension Range:**
- Width: 32–2048 (step: 8)
- Height: 32–2048 (step: 8)
- Invalid values are auto-snapped to the nearest valid step.

**Example:**
```json
{
  "model": "cortex-v1.0",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "mode": "FAST",
    "width": 1024,
    "height": 1024
  }
}
```

---

### Seedance 1.0 Pro Video Gen

`model: seedance-1.0-pro`

Cost: ~$0.09/720p 4s, ~$0.33/720p 15s (pixel×duration based)

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| seed | integer | no | -1 | min: -1, max: 4294967295 | Random seed to control video generation (use -1 for random) |
| width | integer | YES | 1920 | - | Video width in pixels. Use 0 for auto aspect ratio detection (I2V only) |
| height | integer | YES | 1088 | - | Video height in pixels. Use 0 for auto aspect ratio detection (I2V only) |
| prompt | ? | YES | - | maxLen: 5000 | Text description for video generation |
| duration | 2 / 3 / 4 / 5 / 6 / 7 / 8 / 9 / 10 / 11 / 12 | no | 5 | - | Video duration in seconds (2-12 seconds) |
| quantity | integer | no | 1 | min: 1, max: 1 | Number of videos to generate |
| end_frame | ? | no | - | - | Optional final frame for image-to-video generation |
| start_frame | ? | no | - | - | Image to use as first frame for image-to-video generation |
| camera_fixed | boolean | no | false | - | Whether to fix the camera position |
| enable_safety_checker | boolean | no | false | - | Enable content safety filtering |

**Example:**
```json
{
  "model": "seedance-1.0-pro",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "seed": -1,
    "width": 1920,
    "height": 1088,
    "duration": 5,
    "camera_fixed": false,
    "enable_safety_checker": false
  }
}
```

---

### Seedance 1.0 Pro Fast Video Gen

`model: seedance-1.0-pro-fast`

Cost: ~$0.04/720p 4s, ~$0.13/720p 15s (pixel×duration based)

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| seed | integer | no | -1 | min: -1, max: 4294967295 | Random seed to control video generation (use -1 for random) |
| width | integer | YES | 1920 | - | Video width in pixels. Use 0 for auto aspect ratio detection (I2V only) |
| height | integer | YES | 1088 | - | Video height in pixels. Use 0 for auto aspect ratio detection (I2V only) |
| prompt | ? | YES | - | maxLen: 5000 | Text description for video generation |
| duration | 2 / 3 / 4 / 5 / 6 / 7 / 8 / 9 / 10 / 11 / 12 | no | 5 | - | Video duration in seconds (2-12 seconds) |
| quantity | integer | no | 1 | min: 1, max: 1 | Number of videos to generate |
| start_frame | ? | no | - | - | Image to use as first frame for image-to-video generation |
| camera_fixed | boolean | no | false | - | Whether to fix the camera position |
| enable_safety_checker | boolean | no | false | - | Enable content safety filtering |

**Example:**
```json
{
  "model": "seedance-1.0-pro-fast",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "seed": -1,
    "width": 1920,
    "height": 1088,
    "duration": 5,
    "camera_fixed": false,
    "enable_safety_checker": false
  }
}
```

---

### Seedance 2.0 Video Gen

`model: seedance-2.0`

Cost: ~$0.49/720p 4s, ~$1.85/720p 15s (pixel×duration based)

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| seed | integer | no | -1 | min: -1, max: 4294967295 | Random seed to control video generation (use -1 for random) |
| width | integer | no | 1280 | - | Video width in pixels. |
| height | integer | no | 720 | - | Video height in pixels. |
| prompt | ? | YES | - | maxLen: 5000 | Text description for video generation |
| duration | 4 / 5 / 6 / 7 / 8 / 9 / 10 / 11 / 12 / 13 / 14 / 15 | no | 8 | - | Video duration in seconds (4-15 seconds) |
| quantity | integer | no | 1 | min: 1, max: 1 | Number of videos to generate |
| end_frame | ? | no | - | - | Optional final frame for image-to-video generation |
| start_frame | ? | no | - | - | Image to use as first frame for image-to-video generation |
| generate_audio | boolean | no | true | - | Whether to generate audio with the video |
| reference_audio_urls | array | no | - | - | Reference audio files for R2V generation (1-3 files, combined ≤15s). Refer to them as @Audio1, @Audio2, etc. Requires at |
| reference_image_urls | array | no | - | - | Reference images for reference-guided video generation (1-9 images) |
| reference_video_urls | array | no | - | - | Reference videos for video-guided generation (1-3 videos) |
| enable_safety_checker | boolean | no | false | - | Enable content safety filtering |

**Example:**
```json
{
  "model": "seedance-2.0",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "seed": -1,
    "width": 1280,
    "height": 720,
    "duration": 8,
    "generate_audio": true,
    "enable_safety_checker": false
  }
}
```

---

### Seedance 2.0 Fast Video Gen

`model: seedance-2.0-fast`

Cost: ~$0.40/720p 4s, ~$1.48/720p 15s (pixel×duration based)

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| seed | integer | no | -1 | min: -1, max: 4294967295 | Random seed to control video generation (use -1 for random) |
| width | integer | no | 1280 | - | Video width in pixels. |
| height | integer | no | 720 | - | Video height in pixels. |
| prompt | ? | YES | - | maxLen: 5000 | Text description for video generation |
| duration | 4 / 5 / 6 / 7 / 8 / 9 / 10 / 11 / 12 / 13 / 14 / 15 | no | 8 | - | Video duration in seconds (4-15 seconds) |
| quantity | integer | no | 1 | min: 1, max: 1 | Number of videos to generate |
| end_frame | ? | no | - | - | Optional final frame for image-to-video generation |
| start_frame | ? | no | - | - | Image to use as first frame for image-to-video generation |
| generate_audio | boolean | no | true | - | Whether to generate audio with the video |
| reference_audio_urls | array | no | - | - | Reference audio files for R2V generation (1-3 files, combined ≤15s). Refer to them as @Audio1, @Audio2, etc. Requires at |
| reference_image_urls | array | no | - | - | Reference images for reference-guided video generation (1-9 images) |
| reference_video_urls | array | no | - | - | Reference videos for video-guided generation (1-3 videos) |
| enable_safety_checker | boolean | no | false | - | Enable content safety filtering |

**Example:**
```json
{
  "model": "seedance-2.0-fast",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "seed": -1,
    "width": 1280,
    "height": 720,
    "duration": 8,
    "generate_audio": true,
    "enable_safety_checker": false
  }
}
```

---

### Sora 2

`model: sora-2`

Cost: ~$0.04 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| width | 1280 / 720 | no | 1280 | - | - |
| height | 720 / 1280 | no | 720 | - | - |
| prompt | ? | YES | - | - | - |
| duration | 4 / 8 / 12 | no | 4 | - | - |
| quantity | integer | YES | 1 | max: 1 | - |
| start_frame | ? | no | - | - | - |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 16:9 | 1280×720 |
| 9:16 | 720×1280 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "sora-2",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "width": 1280,
    "height": 720,
    "duration": 4
  }
}
```

---

### Sora 2 Pro

`model: sora-2-pro`

Cost: ~$0.12 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| mode | RESOLUTION_720 / RESOLUTION_1080 | no | RESOLUTION_720 | - | - |
| width | 1280 / 720 / 1024 / 1792 | YES | 1280 | - | - |
| height | 720 / 1280 / 1024 / 1792 | YES | 720 | - | - |
| prompt | ? | YES | - | - | - |
| duration | 4 / 8 / 12 | no | 4 | - | - |
| quantity | integer | YES | 1 | max: 1 | - |
| start_frame | ? | no | - | - | - |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 16:9 | 1280×720 |
| 9:16 | 720×1280 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "sora-2-pro",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "mode": "RESOLUTION_720",
    "width": 1280,
    "height": 720,
    "duration": 4
  }
}
```

---

### Veo 3 Fast

`model: veo-3.0-fast-generate-001`

Cost: ~$0.06 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| mode | RESOLUTION_720 / RESOLUTION_1080 | no | RESOLUTION_720 | - | - |
| seed | number | no | - | min: 0, max: 4294967295 | - |
| width | 1280 / 720 / 1920 / 1080 | no | 1280 | - | - |
| height | 720 / 1280 / 1080 / 1920 | no | 720 | - | - |
| prompt | ? | YES | - | - | - |
| duration | 4 / 6 / 8 | no | 8 | - | - |
| quantity | integer | no | 1 | max: 1 | - |
| start_frame | ? | no | - | - | - |
| negative_prompt | ? | no | - | - | - |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 16:9 | 1280×720 |
| 9:16 | 720×1280 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "veo-3.0-fast-generate-001",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "mode": "RESOLUTION_720",
    "width": 1280,
    "height": 720,
    "duration": 8
  }
}
```

---

### Veo 3

`model: veo-3.0-generate-001`

Cost: ~$0.13 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| mode | RESOLUTION_720 / RESOLUTION_1080 | no | RESOLUTION_720 | - | - |
| seed | number | no | - | min: 10, max: 4294967295 | - |
| width | 720 / 1080 / 1280 / 1920 | no | 1280 | - | - |
| height | 720 / 1080 / 1280 / 1920 | no | 720 | - | - |
| prompt | ? | YES | - | - | - |
| duration | 4 / 6 / 8 | no | 8 | - | - |
| quantity | integer | YES | 1 | max: 1 | - |
| start_frame | ? | no | - | - | - |
| negative_prompt | ? | no | - | - | - |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 16:9 | 1280×720 |
| 9:16 | 720×1280 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "veo-3.0-generate-001",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "mode": "RESOLUTION_720",
    "width": 1280,
    "height": 720,
    "duration": 8
  }
}
```

---

### Veo 3.1 Fast

`model: veo-3.1-fast-generate-001`

Cost: ~$0.04 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| mode | RESOLUTION_720 / RESOLUTION_1080 / RESOLUTION_2160 | no | RESOLUTION_720 | - | - |
| seed | number | no | - | min: 0, max: 4294967295 | - |
| audio | ? | no | - | - | - |
| width | 720 / 1080 / 1280 / 1920 / 2160 / 3840 | no | 1280 | - | - |
| height | 720 / 1080 / 1280 / 1920 / 2160 / 3840 | no | 720 | - | - |
| prompt | ? | YES | - | - | - |
| duration | 4 / 6 / 8 | no | 8 | - | - |
| quantity | integer | YES | 1 | max: 1 | - |
| end_frame | ? | no | - | - | - |
| start_frame | ? | no | - | - | - |
| negative_prompt | ? | no | - | - | - |
| motion_has_audio | ? | no | true | - | - |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 16:9 | 1280×720 |
| 9:16 | 720×1280 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "veo-3.1-fast-generate-001",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "mode": "RESOLUTION_720",
    "width": 1280,
    "height": 720,
    "duration": 8,
    "motion_has_audio": true
  }
}
```

---

### Veo 3.1

`model: veo-3.1-generate-001`

Cost: ~$0.08 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| mode | RESOLUTION_720 / RESOLUTION_1080 / RESOLUTION_2160 | no | RESOLUTION_720 | - | - |
| seed | number | no | - | min: 0, max: 4294967295 | - |
| audio | ? | no | - | - | - |
| width | 720 / 1080 / 1280 / 1920 / 2160 / 3840 | no | 1280 | - | - |
| height | 720 / 1080 / 1280 / 1920 / 2160 / 3840 | no | 720 | - | - |
| prompt | ? | YES | - | - | - |
| duration | 4 / 6 / 8 | no | 8 | - | - |
| quantity | integer | YES | 1 | max: 1 | - |
| end_frame | ? | no | - | - | - |
| start_frame | ? | no | - | - | - |
| negative_prompt | ? | no | - | - | - |
| motion_has_audio | ? | no | true | - | - |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 16:9 | 1280×720 |
| 9:16 | 720×1280 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "veo-3.1-generate-001",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "mode": "RESOLUTION_720",
    "width": 1280,
    "height": 720,
    "duration": 8,
    "motion_has_audio": true
  }
}
```

---

### Veo 3.1 Lite

`model: veo-3.1-lite`

Cost: ~$0.02 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| mode | RESOLUTION_720 / RESOLUTION_1080 | no | RESOLUTION_720 | - | - |
| seed | number | no | - | min: 0, max: 4294967295 | - |
| audio | ? | no | - | - | - |
| width | 720 / 1080 / 1280 / 1920 | no | 1280 | - | - |
| height | 720 / 1080 / 1280 / 1920 | no | 720 | - | - |
| prompt | ? | YES | - | - | - |
| duration | 4 / 6 / 8 | no | 8 | - | - |
| quantity | integer | YES | 1 | max: 1 | - |
| end_frame | ? | no | - | - | - |
| start_frame | ? | no | - | - | - |
| negative_prompt | ? | no | - | - | - |
| motion_has_audio | ? | no | true | - | - |

**Supported Resolutions:**

| Label | Width × Height |
|---|---|
| 16:9 | 1280×720 |
| 9:16 | 720×1280 |

Invalid width/height combinations are automatically snapped to the nearest valid pair by aspect ratio and pixel count.

**Example:**
```json
{
  "model": "veo-3.1-lite",
  "type": "video",
  "params": {
    "prompt": "A beautiful scene",
    "mode": "RESOLUTION_720",
    "width": 1280,
    "height": 720,
    "duration": 8,
    "motion_has_audio": true
  }
}
```

---

## 3D Models

### Rodin V2

`model: rodin-v2`

Cost: ~$0.16 per generation

| Parameter | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| seed | integer | no | - | min: 0, max: 65535 | Seed for reproducible generation |
| quality | extra_low / low / medium / high | no | high | - | Generation quality level. Polygon counts depend on mesh_mode: Triangle (extra_low: 2k, low: 20k, medium: 150k, high: 500 |
| ta_pose | boolean | no | false | - | Apply T/A-pose constraint for human character meshes (easier to rig and animate) |
| material | PBR / Shaded / All | no | All | - | Material mode for the generated mesh |
| mesh_mode | Quad / Triangle | no | Triangle | - | Type of faces in the generated mesh. Quad produces quadrilateral faces (smoother surfaces), Triangle produces triangular |
| bounding_box | array | no | - | - | Bounding box dimensions as [width, height, length] integers to constrain the generated model size |
| use_original_alpha | boolean | no | false | - | Preserve original alpha channel from input images |

**3D Output Format:**

The response includes two fields:
- `result` — JPG thumbnail/preview URLs
- `glb_urls` — Downloadable `.glb` 3D model files (same URL with `.jpg` → `.glb`)

Example:
```
result:   https://.../{genId}/rodin-v2_-0.jpg   ← preview image
glb_urls: https://.../{genId}/rodin-v2_-0.glb   ← 3D model download
```

**Example:**
```json
{
  "model": "rodin-v2",
  "type": "3d",
  "params": {
    "quality": "high",
    "ta_pose": false,
    "material": "All",
    "mesh_mode": "Triangle",
    "use_original_alpha": false
  }
}
```

---