> ## Documentation Index > Fetch the complete documentation index at: https://docs.ttapi.io/llms.txt > Use this file to discover all available pages before exploring further. # Reference-to-Video > Reference-to-Video - Generate video based on multiple reference images This is the **HappyHorse Video API** reference-to-video interface, generating videos based on multiple reference images. ## OpenAPI ````yaml openapi/en/happyhorse-reference-to-video.json POST /happyhorse/api/v1/services/aigc/video-generation/video-synthesis openapi: 3.1.0 info: title: HappyHorse Reference-to-Video API version: 1.0.0 description: >- HappyHorse Reference-to-Video API - Generate video based on multiple reference images servers: - url: https://api.ttapi.io security: [] paths: /happyhorse/api/v1/services/aigc/video-generation/video-synthesis: post: description: Reference-to-Video - Generate video based on multiple reference images requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/ReferenceToVideoRequest' responses: '200': $ref: '#/components/responses/HappyHorseSuccessResponse' description: Success Response security: - CustomApiKey: [] components: schemas: ReferenceToVideoRequest: type: object properties: model: type: string description: Model name. Supports happyhorse-1.0-r2v and happyhorse-1.1-r2v. default: happyhorse-1.0-r2v example: happyhorse-1.0-r2v enum: - happyhorse-1.0-r2v - happyhorse-1.1-r2v input: type: object description: Basic input information, including reference images and prompt. properties: prompt: type: string description: >- Text prompt, used to describe the expected elements and visual characteristics in the generated video. Supports any language input, with a maximum length of 5000 non-Chinese characters or 2500 Chinese characters. Exceeding content will be automatically truncated. Reference notation: Use character1, character2 identifiers in prompt to refer to reference images at corresponding positions in the media array, in the same order as the media array. media: type: array description: >- Media list, used to specify reference images. Each element in the array is a media object containing type and url fields. Defines the order of character references in prompt according to array order. The 1st reference_image corresponds to character1, the 2nd to character2, and so on. Reference image count: 1 to 9 images. items: type: object properties: type: type: string description: 'Media type. Fixed value: reference_image.' enum: - reference_image example: reference_image url: type: string description: >- URL address of the image file, must be a publicly accessible URL. Supports HTTP or HTTPS protocol. Image limits: format: JPEG, JPG, PNG, WEBP; resolution: short side not less than 400 pixels, recommend 720P or above clear images, avoid passing images that are too small, blurry, or overly compressed; file size: not exceeding 10MB. required: - type - url required: - prompt - media parameters: type: object description: >- Video generation parameters, such as setting video resolution, aspect ratio, duration, etc. properties: resolution: type: string description: Resolution level of generated video. default: 1080P enum: - 720P - 1080P example: 720P ratio: type: string description: Aspect ratio of generated video. default: '16:9' enum: - '16:9' - '9:16' - '3:4' - '4:3' - '1:1' example: '16:9' duration: type: integer description: >- Duration of generated video in seconds. Value range: integer between 3 and 15. Default: 5. default: 5 minimum: 3 maximum: 15 example: 5 watermark: type: boolean description: >- Whether to add watermark to generated video. Watermark is located at the bottom right corner of video, text is fixed as "Happy Horse". true: default, add watermark; false: no watermark. default: true example: false seed: type: integer description: >- Random seed, value range: [0, 2147483647]. If not specified, system will automatically generate a random seed. To improve reproducibility of generation results, it is recommended to fix the seed value. Note that due to the probabilistic nature of model generation, even with the same seed, cannot guarantee identical results each time. minimum: 0 maximum: 2147483647 example: 12345 required: - model - input responses: HappyHorseSuccessResponse: description: Request successful content: application/json: schema: type: object properties: output: type: object description: Task output information properties: task_status: type: string description: Task status enum: - PENDING - RUNNING - SUCCEEDED - FAILED - CANCELED - UNKNOWN example: PENDING task_id: type: string description: Task ID. Valid for 24 hours. example: 0385dc79-5ff8-4d82-bcb6-xxxxxx request_id: type: string description: >- Request unique identifier. Can be used for request tracing and troubleshooting. example: 4909100c-7b5a-9f92-bfe5-xxxxxx example: output: task_status: PENDING task_id: 0385dc79-5ff8-4d82-bcb6-xxxxxx request_id: 4909100c-7b5a-9f92-bfe5-xxxxxx required: - output - request_id securitySchemes: CustomApiKey: type: apiKey in: header name: TT-API-KEY description: >- Please go to TTAPI Console to get the API key. ````