Canonical integration prompt
Use this first. It is the best default prompt for Cursor, Claude Code, Codex, Windsurf, or your own agent, and it is the version you should hand to users instead of asking them to integrate manually.
You are integrating US Video API. Use only the documented API behavior below and do not invent endpoints, fields, or unsupported workflows.
Base URL:
https://usvideoapi.com/v1
Auth:
- Send Authorization: Bearer <API_KEY>
Supported create endpoint:
- POST /videos
Supported create request fields:
- prompt: string, required
- resolution: one of 480p, 720p, 1080p
- duration: integer, allowed range 4-15
- aspect_ratio: one of 16:9, 9:16, 1:1, 4:3
- image: optional base64 string for image-to-video
- negative_prompt: optional string
- seed: optional integer
Create response:
- id
- status
- model
- resolution
- duration
- price
- estimated_time
Status endpoint:
- GET /videos/{id}
- returns status plus output.video_url when completed
- failed jobs return error
Download:
- use output.video_url or /videos/{id}/download after completion
Error handling:
- 402 insufficient balance
- 429 rate limit / capacity
- 502 upstream or safety failure
- 504 timeout
- read the detail field from error responses
Important constraints:
- Do not assume webhooks, streaming, or auto top-up unless explicitly added by the product later
- Do not assume 2k is currently supported in create requests
- Build polling for async jobs
- Handle failed generations explicitly
- Use visual-first prompts and support retries when safety blocks a prompt
Write production-ready code only.
Paste this into your coding agent before asking it to build an integration. This should be the default recommendation for users who want to move fast.
Backend integration prompt
Build a backend integration for US Video API.
Requirements:
- server-side API key only
- POST /v1/videos to create jobs
- GET /v1/videos/{id} to poll status
- store local job state
- expose clean error handling for 402, 429, 502, 504
- for 502 safety failures, return a retryable message and recommend rewriting the prompt
- do not assume webhooks or streaming
- include typed request/response models
- prefer small polling intervals with backoff
Frontend demo app prompt
Build a frontend demo for US Video API with a small backend proxy. Requirements: - input fields for prompt, resolution, duration, aspect ratio - optional negative prompt - submit create request through backend - poll until completed or failed - show video preview when output.video_url exists - if failed, display the API detail string clearly - include a retry button - do not expose the API key in the browser
Image-to-video prompt
Integrate image-to-video for US Video API. Requirements: - accept an uploaded image - convert it to base64 - send the base64 string in the image field to POST /v1/videos - include prompt, resolution, duration, and aspect_ratio - include negative_prompt support - validate file size and type before upload - keep prompts product-focused and visual-first - handle 502 safety failures without infinite retries
Enterprise account prompt
Build account management UI for US Video API enterprise users. Requirements: - support shared accounts with roles - show account balance, transactions, API keys, jobs, and members - treat owner, billing_admin, developer, and viewer differently - do not invent SSO, SCIM, or multi-workspace features - keep flows aligned with current shared-account model - expose billing and access controls clearly - show failed job errors and safety-related rejections in the job history
Why this matters
Without a prompt pack
- Agents invent fields that do not exist
- They assume unsupported workflows
- They skip safety and retry logic
With a prompt pack
- Integrations start from the real API shape
- Error handling is correct from day one
- Teams move faster with less support overhead
Related resources: Use the API docs for endpoint details, the Prompt Safety Guide for moderation-safe prompt structure, and the enterprise guide for shared-account workflows.