What your agent can do
- Generate videos from text, an image, reference assets or an existing clip, from a single request in plain language
- Pick the right model for the job: speed, cost, resolution, duration and audio
- Follow the task from
queuedtocompletedand hand you the video URL - Write your integration: API client, retries with idempotency keys, a webhook endpoint with signature verification
- Manage credits: check your balance before generating and explain what each task was charged
Install
Run this in your project:Try it
Ask your agent, for example:Generate a 5-second 480p video of a paper boat floating down a rain gutter with Five API.
Animate https://example.com/my-product.png into a slow 360-degree orbit, 1080p, with Wan 3.0 Prime Spark.
Add a Five API webhook endpoint to this Express app that verifies the signature and stores completed videos in S3.
Write a Python worker that creates Five API tasks from our job queue, retries safely and stays within the concurrency limit.The agent chooses a model, creates the task, waits for the result and gives you the video URL, or writes the code into your project.
Install manually
The skill is a single Markdown file, published at /skill.md. To install it by hand, for example for Claude Code:Prerequisites
- A Five API account and an API key from fiveapi.ai (API Keys). See Authentication.
- Credits on your balance. See Credits and pricing.
- Public HTTPS URLs for any input image, video or audio. Five API has no upload endpoint; see Input files.
Also useful
Docs for AI assistants
Connect these docs to your assistant through MCP, llms.txt or Markdown.
API reference
Every endpoint, with an interactive playground.