> ## Documentation Index > Fetch the complete documentation index at: https://docs.aiohub.org/llms.txt > Use this file to discover all available pages before exploring further. # Create video ยท Gemini API > Submit a Seedance video task using Gemini API, with examples, input limits and full parameters. Call Seedance through the Gemini API long-running operation format. Poll with GET; successful results appear in `response.generateVideoResponse.generatedSamples`. For a client using Vertex AI paths, follow the [Vertex AI endpoint](/en/api/video-vertex). ## Setup Use `https://api.aiohub.org` and authenticate with an AIOHub API key. Replace `` with your full `sk-` key and confirm its group can access the selected Seedance model. ## Example request ```bash theme={"system"} curl "https://api.aiohub.org/v1beta/models/doubao-seedance-2-5-260628:predictLongRunning" \ -H "Authorization: Bearer " \ -H "Content-Type: application/json" \ -d '{ "instances": [{ "prompt": "A blue ceramic cup on a wooden table in daylight." }], "parameters": { "durationSeconds": 4, "aspectRatio": "16:9", "resolution": "720p", "generateAudio": true, "sampleCount": 1 } }' ``` The response includes the fields below. Save the full `name`, including the model path and `operations/`, rather than only its final ID segment. ```json theme={"system"} { "name": "models/doubao-seedance-2-5-260628/operations/task_example", "done": false } ``` After saving the task identifier, [poll the task](/en/api/video-gemini-get). ## Inputs and parameters Omitted controls default to 5 seconds, 720p, `adaptive` aspect ratio and audio enabled. First/last frames and reference images are separate input modes; do not combine them in the same `instances[0]`. Example image object: ```json theme={"system"} {"bytesBase64Encoded": "", "mimeType": "image/png"} ``` See [Video models and pricing](/en/api/video-models) for model limits. Arbitrary `gs://` inputs, video extension and the full Veo parameter set are not supported. Use [Ark / Seedance](/en/api/video-ark) for reference audio, reference video or video editing. Install `google-genai` and set `AIOHUB_API_KEY` in your environment. This example uses Gemini API mode. The base URL excludes `/v1beta`; set the version with `api_version`. ```python Python theme={"system"} import os import time from google import genai from google.genai import types client = genai.Client( api_key=os.environ["AIOHUB_API_KEY"], http_options=types.HttpOptions(base_url="https://api.aiohub.org", api_version="v1beta"), ) op = client.models.generate_videos( model="doubao-seedance-2-0-260128-libtv", source=types.GenerateVideosSource(prompt="A blue ceramic cup on a wooden table."), config=types.GenerateVideosConfig( number_of_videos=1, duration_seconds=4, aspect_ratio="16:9", resolution="720p" ), ) while not op.done: time.sleep(15) op = client.operations.get(op) if op.error: raise RuntimeError(str(op.error)) video = op.response.generated_videos[0].video with open("video.mp4", "wb") as f: f.write(client.files.download(file=video)) ``` If creation times out without a task identifier, check the [usage log](https://api.aiohub.org/console/log) before resubmitting to avoid duplicate generation and charges. Correct parameters after a 400 and check the API key after a 401. See [Authentication](/en/api/authentication) for access and quota requirements. ## OpenAPI ````yaml openapi/relay.en.json POST /v1beta/models/{model}:predictLongRunning openapi: 3.0.1 info: title: AIOHub API description: AIOHub public API version: 1.0.0 servers: - url: https://api.aiohub.org security: - BearerAuth: [] tags: - name: List models - name: OpenAI Chat - name: OpenAI Responses - name: Image generation - name: OpenAI Images - name: Video generation - name: Claude Messages - name: Gemini API - name: OpenAI Embeddings - name: Text completions - name: OpenAI Audio - name: Realtime API paths: /v1beta/models/{model}:predictLongRunning: post: tags: - Video generation summary: Create Gemini video task description: > This page covers the Seedance Gemini video format. See [Models and pricing](/en/api/video-models) for model limits. Creation returns name and done: false. Save the full name and poll every 10 to 15 seconds. HTTP 200 does not mean the video is ready. operationId: createSeedanceGeminiVideo parameters: - name: model in: path required: true description: Full Seedance model name. schema: type: string example: doubao-seedance-2-5-260628 example: doubao-seedance-2-5-260628 requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/SeedanceGoogleVideoRequest' examples: text: summary: Text to video value: instances: - prompt: A blue ceramic cup on a wooden table in daylight. parameters: durationSeconds: 4 aspectRatio: '16:9' resolution: 720p generateAudio: true sampleCount: 1 responses: '200': description: Task accepted. content: application/json: schema: $ref: '#/components/schemas/SeedanceGeminiOperation' example: name: models/doubao-seedance-2-5-260628/operations/task_example done: false '400': description: >- Invalid fields, types or model capability. Correct the request using code and message before retrying. content: application/json: schema: $ref: '#/components/schemas/SeedanceTaskError' '401': description: Missing or invalid API key. '429': description: Too many requests or capacity exhausted; wait before retrying. '503': description: >- Model unavailable in the selected group, or service temporarily unavailable. security: - BearerAuth: [] components: schemas: SeedanceGoogleVideoRequest: type: object properties: instances: type: array minItems: 1 maxItems: 1 description: >- Send exactly one instance. Text generation requires prompt; use image for a first frame. items: type: object properties: prompt: type: string description: Text prompt, required for text-only generation. image: $ref: '#/components/schemas/SeedanceGoogleImage' lastFrame: $ref: '#/components/schemas/SeedanceGoogleImage' referenceImages: type: array minItems: 1 maxItems: 30 items: $ref: '#/components/schemas/SeedanceGoogleReferenceImage' description: >- Up to 30 images for 2.5 or 9 for the 2.0 family. Cannot be combined with image or lastFrame. parameters: type: object properties: durationSeconds: type: integer minimum: 4 maximum: 30 default: 5 description: 'Whole seconds: 4 to 30 for 2.5, 4 to 15 for the 2.0 family.' aspectRatio: type: string description: >- Defaults to adaptive. Seedance 2.5 first/last-frame input requires adaptive. enum: - adaptive - '16:9' - '4:3' - '1:1' - '3:4' - '9:16' - '21:9' default: adaptive resolution: type: string description: >- Model-dependent. 4k is available only on standard 2.0; fast/mini do not support 1080p. enum: - 480p - 720p - 1080p - 4k default: 720p generateAudio: type: boolean default: true description: Whether to generate audio. sampleCount: type: integer enum: - 1 default: 1 description: Each task generates one video. numberOfVideos: type: integer enum: - 1 description: Alias for sampleCount; must be 1 when supplied. required: - instances description: >- This schema covers Seedance. lastFrame requires image; reference images cannot be combined with first/last frames. Video extension, masks, seed, negativePrompt and other unlisted controls are unsupported. SeedanceGeminiOperation: type: object properties: name: type: string description: Full operation name. Save the returned value for polling. done: type: boolean description: >- False means pending. True means finished, possibly with a failure; check error. response: type: object properties: '@type': type: string description: Response type. generateVideoResponse: type: object properties: generatedSamples: type: array items: type: object properties: video: type: object properties: uri: type: string description: >- Full HTTPS file download URL returned on success; authentication is required. format: uri required: - uri required: - video error: type: object properties: message: type: string description: Generation failure reason. required: - message required: - name - done SeedanceTaskError: type: object description: Error response of the video task endpoints. properties: code: type: string description: Error code example: invalid_duration message: type: string description: Error message example: duration must be between 4 and 30 seconds for this model SeedanceGoogleImage: type: object properties: bytesBase64Encoded: type: string description: Base64 image bytes without a data URL prefix. mimeType: type: string description: Image MIME type, such as image/png or image/jpeg. example: image/png required: - bytesBase64Encoded - mimeType SeedanceGoogleReferenceImage: type: object properties: image: $ref: '#/components/schemas/SeedanceGoogleImage' referenceType: type: string description: Use asset. Do not combine references with first/last-frame inputs. enum: - asset default: asset required: - image - referenceType securitySchemes: BearerAuth: type: http scheme: bearer description: API key (sk-xxx) ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.