### GET /unified-video/models

Source: https://midsync.dev/zh-TW/developers/docs/video-models-api

取得影片模型的 runtime metadata，包含可用模型清單、欄位定義、選項與限制。

```APIDOC
## GET /unified-video/models

### Description
取得影片模型的 runtime metadata，讓前端知道有哪些模型、欄位、選項與限制可以安全渲染。此 API 應作為影片表單渲染前的 source of truth。

### Method
GET

### Endpoint
/unified-video/models

### Parameters
#### Auth
- **Authorization** (header) - Required - 使用產品登入狀態，或 Bearer <midsync_api_key>

### Response
#### Success Response (200)
- **models** (array) - 可用的影片模型清單
- **modelParams** (object) - 欄位定義、可選 enum、預設值與模型限制
- **pricingHints** (object) - 目前價格提示或價格範例

### Request Example
```javascript
const models = await fetch("/unified-video/models", {
  credentials: "include"
}).then((resp) => resp.json());
```
```

--------------------------------

### GET /v1/usage/logs

Source: https://midsync.dev/zh-TW/developers/docs/usage-logs-api

Retrieves detailed usage logs, suitable for generating billing tables and detail pages. Requires 'api.read' permission.

```APIDOC
## GET /v1/usage/logs

### Description
Retrieves granular usage history, ideal for detailed billing tables and historical tracking. This endpoint is designed for scenarios where clients need to trace every expenditure.

### Method
GET

### Endpoint
/v1/usage/logs

### Parameters
#### Query Parameters
- **page** (integer) - Optional - Used for pagination if supported by the backend.
- **limit** (integer) - Optional - Used for pagination if supported by the backend.
- **date_range** (string) - Optional - Specifies the historical date range for the logs (e.g., 'last_7_days', '2023-01-01_2023-12-31').
- **model** (string) - Optional - Filters logs for a specific model if supported by the backend.

### Request Example
(No request body for GET requests)

### Response
#### Success Response (200)
- **logs** (array) - An array of usage log objects.
  - **model** (string) - The model used for this request.
  - **pointsCharged** (integer) - The number of points deducted for this request.
  - **usage** (object) - Contains token or other usage information if relevant.
  - **createdAt** (string) - The timestamp when the request was created.

#### Response Example
```json
{
  "logs": [
    {
      "model": "grok-video",
      "pointsCharged": 100,
      "usage": {
        "duration": 10,
        "resolution": "720p"
      },
      "createdAt": "2023-10-27T10:00:00Z"
    }
  ]
}
```
```

--------------------------------

### GET /v1/usage/summary

Source: https://midsync.dev/zh-TW/developers/docs/usage-summary-api

Retrieves aggregated usage data, suitable for dashboard and billing summary cards.

```APIDOC
## GET /v1/usage/summary

### Description
This API is suitable for high-level summary cards. It complements the usage logs API by providing an overview, while the logs API offers detailed breakdowns.

### Method
GET

### Endpoint
/v1/usage/summary

### Parameters
#### Query Parameters
- **date_range** (string) - Optional - Specifies the summary period if supported by the backend.
- **groupBy** (string) - Optional - Aggregates data by dimensions such as models, if supported by the backend.

### Response
#### Success Response (200)
- **totalPointsCharged** (integer) - Total points charged within the specified period.
- **totalRequests** (integer) - Total requests made within the specified period.
- **totalInputTokens** (integer) - Total input tokens within the specified period, if applicable.
- **totalOutputTokens** (integer) - Total output tokens within the specified period, if applicable.
- **byModel** (object) - Summary data for each model.
  - **model_name** (object) - Details for a specific model.
    - **pointsCharged** (integer) - Points charged for this model.
    - **requests** (integer) - Requests made for this model.
    - **inputTokens** (integer) - Input tokens for this model, if applicable.
    - **outputTokens** (integer) - Output tokens for this model, if applicable.

### Request Example
```json
{
  "example": "GET /v1/usage/summary?date_range=last_7_days&groupBy=model"
}
```

### Response Example
```json
{
  "totalPointsCharged": 15000,
  "totalRequests": 1000,
  "totalInputTokens": 500000,
  "totalOutputTokens": 750000,
  "byModel": {
    "gpt-4": {
      "pointsCharged": 10000,
      "requests": 600,
      "inputTokens": 300000,
      "outputTokens": 450000
    },
    "gpt-3.5-turbo": {
      "pointsCharged": 5000,
      "requests": 400,
      "inputTokens": 200000,
      "outputTokens": 300000
    }
  }
}
```
```

--------------------------------

### Get Video Quote Estimate

Source: https://midsync.dev/zh-TW/developers/docs/video-quote-api

Use this snippet to make a POST request to the /unified-video/quote endpoint to obtain an estimated point cost for video generation. Ensure the request body matches the /generate endpoint's requirements. Re-quote if model, duration, resolution, quality, audio, or mode changes.

```javascript
const quote = await fetch("/unified-video/quote", {
  method: "POST",
  credentials: "include",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    model: "seedance-1.5-pro",
    prompt: "一隻狗在明亮攝影棚裡奔跑的電影感鏡頭",
    duration: 8,
    resolution: "720p"
  })
}).then((resp) => resp.json());
```

--------------------------------

### Fetch Video Models Metadata

Source: https://midsync.dev/zh-TW/developers/docs/video-models-api

Call this API before rendering video forms to get a source of truth for available models, field enums, default values, and constraints. Ensure credentials are included for authentication.

```typescript
const models = await fetch("/unified-video/models", {
  credentials: "include"
}).then((resp) => resp.json());
```

--------------------------------

### GET /unified-video/:taskId/status

Source: https://midsync.dev/zh-TW/developers/docs/video-status-api

Retrieves the current status of a video generation task using the taskId.

```APIDOC
## GET /unified-video/:taskId/status

### Description
Retrieves the current status of a video generation task. This endpoint is intended for manual queries, fallback checks, or recovery processes rather than continuous polling.

### Method
GET

### Endpoint
/unified-video/:taskId/status

### Parameters
#### Path Parameters
- **taskId** (string) - Required - The unique identifier of the task returned from the /unified-video/generate endpoint.

### Response
#### Success Response (200)
- **success** (boolean) - Indicates if the query was successful.
- **taskId** (string) - The ID of the queried task.
- **status** (string) - The current status of the task.
- **output / assets** (object) - The generated result if the task is completed.
- **error** (string) - Error information if the task failed.
```

--------------------------------

### Create Video Generation Task

Source: https://midsync.dev/zh-TW/developers/docs/video-generate-api

Use this cURL command to send a POST request to the /unified-video/generate endpoint to create a video generation task. Ensure you replace `<midsync_api_key>` with your actual API key. The request body specifies the model, prompt, duration, and resolution for the video.

```bash
curl --request POST \
  --url https://api-dev.midsync.ai/unified-video/generate \
  --header "Authorization: Bearer <midsync_api_key>" \
  --header "Content-Type: application/json" \
  --data '{ \
    "model": "seedance-1.5-pro", \
    "prompt": "黃昏海邊的電影感鏡頭，海浪輕拍岸邊", \
    "duration": 8, \
    "resolution": "720p" \
  }'
```

--------------------------------

### POST /unified-video/generate

Source: https://midsync.dev/zh-TW/developers/docs/model-runway

Generates a video using the Runway model based on text prompts or image inputs.

```APIDOC
## POST /unified-video/generate

### Description
Generates a video using the Runway model. Supports text-to-video and image-to-video generation.

### Method
POST

### Endpoint
/unified-video/generate

### Parameters
#### Request Body
- **model** (string) - Required - Must be set to "runway".
- **prompt** (string) - Required - Description of the video content.
- **images** (array) - Optional - Maximum 1 image for image-guided generation.
- **aspectRatio** (string) - Optional - Required if no image is provided.
- **duration** (integer) - Optional - Video length in seconds.
- **quality** (string) - Optional - Output quality (e.g., 720p, 1080p).
- **watermark** (boolean) - Optional - Toggle for video watermark.
- **projectId** (string) - Optional - Internal tracking ID.

### Request Example
{
  "model": "runway",
  "prompt": "A cinematic shot of a mountain landscape",
  "aspectRatio": "16:9",
  "duration": 5,
  "quality": "720p"
}
```

--------------------------------

### POST /unified-video/generate

Source: https://midsync.dev/zh-TW/developers/docs/model-kling-25

Endpoint to trigger video generation using the Kling 2.5 Turbo model.

```APIDOC
## POST /unified-video/generate

### Description
Generates a video based on text prompts or image inputs using the kling-2.5-turbo model.

### Method
POST

### Endpoint
/unified-video/generate

### Parameters
#### Request Body
- **model** (string) - Required - Must be "kling-2.5-turbo".
- **prompt** (string) - Required - Description of the video content.
- **images** (array) - Optional - List of image URLs (max 2) for image-to-video generation.
- **aspectRatio** (string) - Optional - Output aspect ratio (e.g., "16:9", "9:16").
- **duration** (integer) - Optional - Video duration in seconds (5 or 10).
- **negativePrompt** (string) - Optional - Content to avoid in the video.
- **cfgScale** (float) - Optional - Adherence to prompt (0 to 1).
- **projectId** (string) - Optional - Internal tracking ID.

### Request Example
{
  "model": "kling-2.5-turbo",
  "prompt": "A cinematic shot of a futuristic city",
  "duration": 5,
  "aspectRatio": "16:9"
}
```

--------------------------------

### POST /unified-video/generate

Source: https://midsync.dev/zh-TW/developers/docs/video-generate-api

Creates a video generation task after client acceptance of settings and pricing. This endpoint is suitable for teams implementing a submission flow, task creation, success status tracking, and final point deduction.

```APIDOC
## POST /unified-video/generate

### Description
Creates a video generation task. This endpoint should be used after a client has accepted the settings and pricing. It is designed for teams that need to implement a submission flow, task creation, track success status, and manage final point deductions. The primary method for receiving completion results is via callback.

### Method
POST

### Endpoint
https://api-dev.midsync.ai/unified-video/generate

### Parameters
#### Request Body
- **model** (string) - Required - The ID of the selected video model.
- **prompt** (string) - Required - The main generation prompt.
- **duration** (integer) - Optional (may be required depending on the model) - The desired duration of the video.
- **resolution** (string) - Optional (may be required depending on the model) - The desired resolution of the video (e.g., "720p").
- **quality** (string) - Optional - Used by quality grading models.
- **aspectRatio** (string) - Optional - Must be an aspect ratio supported by the model.
- **mode** (string) - Optional - Used for multi-mode models.
- **generateAudio** (boolean) - Optional - Send only if the model supports audio generation.
- **images** (array of strings) - Optional - Provide image URLs for image-to-video models.
- **referenceVideoUrls** (array of strings) - Optional - Provide URLs for reference videos if the model supports it.

### Request Example
```json
{
  "model": "seedance-1.5-pro",
  "prompt": "黃昏海邊的電影感鏡頭，海浪輕拍岸邊",
  "duration": 8,
  "resolution": "720p"
}
```

### Response
#### Success Response (200)
- **success** (boolean) - Indicates if the task creation was successful.
- **taskId** (string) - The internal task ID, used for subsequent status checks.
- **providerTaskId** (string) - The task ID from the upstream provider.
- **model** (string) - The ID of the model used for generation.
- **providerModel** (string) - The actual model name used by the provider.
- **pointsCharged** (integer) - The number of points actually deducted for this generation.
- **status** (string) - The initial status of the task (e.g., "submitted").
- **estimatedTime** (string) / **estimatedTimeSec** (integer) - Estimated processing time, if provided by the backend.

#### Response Example
```json
{
  "success": true,
  "taskId": "task_12345abc",
  "providerTaskId": "provider_xyz789",
  "model": "seedance-1.5-pro",
  "providerModel": "sdxl-video-v1",
  "pointsCharged": 10,
  "status": "submitted",
  "estimatedTimeSec": 120
}
```

### Frontend Usage Notes
- Disable the primary button upon submission.
- Display `pointsCharged` directly after successful creation.
- The primary completion flow should be handled via callback. The `/unified-video/:taskId/status` endpoint should be considered a fallback, manual query, or recovery mechanism.
```

--------------------------------

### Sample Request Body for Price Estimation

Source: https://midsync.dev/zh-TW/pricing-estimator

A JSON object representing the configuration for a video generation request. This structure dynamically updates based on user-selected parameters.

```json
{
  "model": "seedance-1.5-pro",
  "prompt": "黃昏海邊的電影感鏡頭，光線柔和，海浪慢慢推進",
  "duration": 8,
  "resolution": "720p",
  "aspectRatio": "16:9"
}
```

--------------------------------

### POST /unified-video/quote

Source: https://midsync.dev/zh-TW/developers/docs/video-quote-api

Estimates the point cost for a video generation request based on the provided parameters.

```APIDOC
## POST /unified-video/quote

### Description
Pre-calculates the point cost for a video generation request. The request body should match the structure used for the generation endpoint.

### Method
POST

### Endpoint
/unified-video/quote

### Parameters
#### Request Body
- **model** (string) - Required - Video model ID (e.g., seedance-1.5-pro)
- **prompt** (string) - Required - Main generation prompt
- **duration** (number) - Required (model dependent) - Video duration
- **resolution** (string) - Required (model dependent) - Video resolution
- **quality** (string) - Optional - Model-specific quality setting
- **aspectRatio** (string) - Optional - Supported aspect ratio
- **mode** (string) - Optional - Model mode
- **generateAudio** (boolean) - Optional - Whether to include audio pricing
- **images** (array) - Optional - Array of image URLs for image-to-video
- **referenceVideoUrls** (array) - Optional - Array of reference video URLs

### Request Example
{
  "model": "seedance-1.5-pro",
  "prompt": "一隻狗在明亮攝影棚裡奔跑的電影感鏡頭",
  "duration": 8,
  "resolution": "720p"
}

### Response
#### Success Response (200)
- **success** (boolean) - Whether the estimation was successful
- **model** (string) - The model ID used for estimation
- **pointsCharged** (number) - Estimated point cost
- **error** (string) - Error message if the combination is invalid
```

--------------------------------

### Kling 2.6 T2V Generation

Source: https://midsync.dev/zh-TW/developers/docs/model-kling-26-t2v

Endpoint documentation for generating videos from text using the kling-2.6-t2v model.

```APIDOC
## POST /unified-video/generate

### Description
Generates a video based on a text prompt using the Kling 2.6 T2V model. This endpoint requires the 'api.generate' permission.

### Method
POST

### Endpoint
/unified-video/generate

### Parameters
#### Request Body
- **model** (string) - Required - Must be 'kling-2.6-t2v'.
- **prompt** (string) - Required - Description of video content, action, style, and camera movement.
- **aspectRatio** (string) - Optional - Controls output aspect ratio.
- **duration** (integer) - Optional - Video length in seconds (supports 5 or 10).
- **generateAudio** (boolean) - Optional - Whether to generate audio. Enabling this increases the cost.
- **negativePrompt** (string) - Optional - Features to exclude from the video.
- **cfgScale** (float) - Optional - Adherence to prompt (0.0 to 1.0).
- **cameraControl.cameraMotion** (string) - Optional - Camera movement type (e.g., pan, tilt, zoom).
- **cameraControl.cameraSpeed** (float) - Optional - Camera speed (0.1 to 2.0).
- **projectId** (string) - Optional - Internal tracking ID.

### Request Example
{
  "model": "kling-2.6-t2v",
  "prompt": "A cinematic shot of a futuristic city",
  "duration": 5,
  "generateAudio": false,
  "cfgScale": 0.5
}

### Response
#### Success Response (200)
- **status** (string) - Generation status.
- **jobId** (string) - Unique identifier for the generation task.
```

--------------------------------

### 影音生成任務請求範例

Source: https://midsync.dev/zh-TW/developers/docs/video-status-api

用於發起影音生成任務的 JSON 請求格式。

```json
{
  "model": "hailuo-2.3-standard",
  "prompt": "讓人像照片產生自然的呼吸與眨眼動作",
  "images": [
    "https://example.com/portrait.png"
  ],
  "duration": 10,
  "resolution": "768P"
}
```

--------------------------------

### Kling 3.0 Model Configuration

Source: https://midsync.dev/zh-TW/developers/docs/model-kling-30

Details on parameters and rules for the Kling 3.0 model generation.

```APIDOC
## POST /unified-video/generate

### Description
Generates video content using the Kling 3.0 model. Supports both single-shot and multi-shot modes.

### Method
POST

### Endpoint
/unified-video/generate

### Request Body
- **model** (string) - Required - Must be "kling-3.0"
- **prompt** (string) - Required (if single-shot) - Main prompt for single-shot mode
- **images** (array) - Optional - Reference images
- **aspectRatio** (string) - Optional - "16:9", "9:16", or "1:1"
- **duration** (integer) - Optional - Video length
- **generateAudio** (boolean) - Optional - Whether to generate audio
- **mode** (string) - Optional - "std" or "pro"
- **multiShots** (boolean) - Optional - Set to true for multi-shot mode
- **multiPrompt** (array) - Optional - Required if multiShots is true. Contains prompt and duration (1-12) for each shot
- **klingElements** (array) - Optional - Element references (max 3 groups, 2-4 images per group)
- **projectId** (string) - Optional - Internal tracking ID

### Request Example
{
  "model": "kling-3.0",
  "prompt": "A cinematic shot of a futuristic city",
  "mode": "std",
  "aspectRatio": "16:9"
}
```

--------------------------------

### seedance-1.5-pro Video Generation API

Source: https://midsync.dev/zh-TW/developers/docs/model-seedance-15

This API endpoint allows for video generation using the seedance-1.5-pro model. It supports both text-to-video and image-to-video generation with various customization options.

```APIDOC
## POST /unified-video/generate

### Description
Generates video content using the seedance-1.5-pro model. Supports text-to-video and image-to-video generation.

### Method
POST

### Endpoint
/unified-video/generate

### Parameters
#### Request Body
- **model** (string) - Required - Specifies the video model to use. Must be 'seedance-1.5-pro'.
- **prompt** (string) - Required - The main text prompt describing the desired video content, scene, style, and action.
- **images** (array[string]) - Optional - An array of image URLs (0 to 2) to guide video generation. 0 images for text-to-video, 1-2 images for image-guided video.
- **aspectRatio** (string) - Optional - The desired aspect ratio of the output video (e.g., '16:9', '9:16'). Supported values: '1:1', '4:3', '3:4', '16:9', '9:16', '21:9', '2:3', '3:2'.
- **resolution** (string) - Optional - The desired resolution of the output video. Supported values: '480p', '720p', '1080p'.
- **duration** (integer) - Optional - The desired duration of the video in seconds. Supported values: 4, 8, 12.
- **generateAudio** (boolean) - Optional - Whether to generate audio along with the video. Defaults to false. Enabling this may affect pricing.
- **fixedLens** (boolean) - Optional - Whether to maintain a more fixed camera lens effect. Use only when necessary.
- **nsfwChecker** (boolean) - Optional - Whether to enable content safety checks. This is a platform control field.
- **projectId** (string) - Optional - An internal tracking field for accounting, querying, or project data integration.

### Request Example
```json
{
  "model": "seedance-1.5-pro",
  "prompt": "A majestic dragon flying over a medieval castle at sunset, cinematic style.",
  "aspectRatio": "16:9",
  "resolution": "1080p",
  "duration": 12,
  "generateAudio": true
}
```

### Response
#### Success Response (200)
- **videoId** (string) - The unique identifier for the generated video.
- **status** (string) - The current status of the video generation process.

#### Response Example
```json
{
  "videoId": "vid_abc123xyz",
  "status": "processing"
}
```
```

--------------------------------

### 呼叫用量紀錄 API 請求範例

Source: https://midsync.dev/zh-TW/developers/docs/usage-logs-api

此 JSON 範例展示了請求模型生成任務時的參數結構，包含模型選擇、提示詞與解析度設定。

```json
{
  "model": "grok-video",
  "prompt": "一隻金毛犬在明亮攝影棚中向鏡頭跑來",
  "duration": 10,
  "resolution": "720p",
  "aspectRatio": "16:9"
}
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.