Vidance
ModelsToolsShowcaseAPIPricingBlog
Loading...Start for free

API documentation

QuickstartAuthenticationModel catalogGeneration tasksHow you're chargedStatus and resultsErrors

Keep API keys on your server. Never expose a long-lived key in browser code or client logs.

Get started with the Vidance API

Use one predictable interface to access leading generation models.

Live request builder

Generate a request for any available model

Choose a live catalog entry to see a schema-based request with the exact identity and supported input shape.

Server-side request
Loading catalog…

Values are generated from the live input schema. Replace placeholders before sending.

Quickstart

Create an API key and send your first generation request.

MethodGET /developer/v1/models
MethodGET /developer/v1/models/form-schema
MethodPOST /developer/v1/generations
  1. 1Create and securely store a Developer API key
  2. 2Resolve a catalog identity and submit the generation task
  3. 3Poll task status and refresh temporary output URLs

Authentication

Send Authorization: Bearer lpk_... from your server. Keys have query and generate scopes, are shown once, and must never appear in browser code.

HeaderAuthorization: Bearer $VIDANCE_API_KEY
HeaderIdempotency-Key: $IDEMPOTENCY_KEY
Scopesgenerate · query

Model catalog

Resolve the model ID, capability, and schema version from the business model catalog instead of hardcoding provider assumptions.

MethodGET /developer/v1/models?limit={limit}&offset={offset}
MethodGET /developer/v1/models/form-schema
Returnsmodel_id · capability · schema_version · input_schema

Generation tasks

Submit the complete catalog identity and schema-valid input with a stable Idempotency-Key header. The generate scope is required.

MethodPOST /developer/v1/generations
Bodymodel_id · capability · schema_version · input
Returns202 · task_id · status

How you're charged

Developer API charges use your prepaid USD wallet with hold-and-settle billing tied to each generation task.

MethodPOST /developer/v1/billing/estimate
Returnsheld USD for the selected model

1.A hold is placed first

When you submit a generation, Vidance places a hold for the estimated maximum charge in your API wallet.

2.The actual price is charged when the job succeeds

After the job completes successfully, the final metered amount is charged and any unused hold is released.

3.You never pay more than the hold

If metering finishes below the hold, only the result amount is charged and the remainder returns to available balance.

Example: a $6.00 hold with a $4.50 result charges $4.50 and returns the difference to available balance.

Status and results

Use the query scope to poll task status, list stable output indexes, and refresh temporary content URLs. Cancellation is not supported.

MethodGET /developer/v1/generations?limit={limit}&offset={offset}
MethodGET /developer/v1/generations/{task_id}
MethodGET /developer/v1/generations/{task_id}/outputs/{output_index}/content
Returnsurl · url_expires_at

Errors

Handle structured error codes and request identifiers without exposing sensitive request data.

Envelope{ code, message, request_id }
Status400 Bad Request · 401 Unauthorized
Status404 Not Found · 409 Conflict · 410 Gone
RetryIdempotency-Key
Vidance

Vidance is an all-in-one AI video and image generation platform that brings together leading models such as Seedance, Kling, Veo, Sora, and Hailuo. Creators, businesses, and developers can generate videos from text or images, create AI visuals, enhance content with practical tools, or access multiple models through a unified API.

Features

Creative studioAI toolsModel libraryAPI platform

Video models

Seedance 2.0Kling 3.0HappyHorse 1.1Wan 3.0More models

Developers / Business

API documentationResource centerPricingAbout usPrivacy policy

© 2026 Vidance.ai. All rights reserved.

Follow Vidance on XJoin the Vidance Discord communityWatch Vidance on YouTubeView Vidance on GitHub