/v1 backward compatible. New models, new optional request fields, new response fields and new error codes are added without a version change, so make your client ignore fields and codes it doesn’t know. A breaking change would ship under a new path prefix (/v2), and /v1 would keep working for at least 12 months after that.
Five API v1
The first public release of the Five API.Models- Video Fast 1.0: text to video, image to video and reference to video at 480p and 768p
- Wan 3.0 Spark: text to video, image to video and reference to video up to 1080p and 30 seconds, with synchronized audio
- Wan 3.0 Prime Spark: text to video and image to video, Wan 3.0 quality with a faster turnaround
- Video Fast Extend 1.1: continue an existing clip by up to 30 seconds, with matched audio
- One create endpoint per model,
POST /v1/{brand}/{version}/{task}, with published input schemas (GET /v1/models) - Tasks: get, list with filters and cursor paging, and cancel with an immediate credit release
- Account balance, plan limits and daily usage
- Signed webhooks (Standard Webhooks) for
task.completed,task.failedandtask.cancelled, with automatic retries - Safe retries with
Idempotency-Key, and a stable error format with a request id on every response
- An agent skill and a docs MCP server for AI coding agents