FAQ

What is Pipevideo?

Pipevideo is a generative video gateway. You send a prompt (and optional reference media) via an OpenAI Responses-compatible API; Pipevideo orchestrates an LLM, renders with HyperFrames, and returns a video URL.

How is pricing calculated?

HyperFrames generations bill the orchestration LLM by token usage (input and output tokens, per-million rates listed on Engines). Credits are deducted after orchestration completes. There is no flat per-video fee — cost scales with prompt complexity and model choice.

What API shape does Pipevideo use?

Pipevideo exposes an OpenAI Responses-compatible /v1/responses endpoint. Requests use input (and optional instructions); completed responses include output_video_url and typed output items with video content.

Do I need to poll for results?

Generations are asynchronous. You can poll GET /v1/responses/{id} every few seconds, or configure webhooks for push notifications when generations finish.

What are the rate limits?

Authenticated requests are limited to 100 requests per minute per API key. When exceeded, the API returns 429 Too Many Requests with a retry_after value in seconds.

Can I use the OpenAI SDK?

Yes. Point the OpenAI client at https://api.pipevideo.co/v1 with your Pipevideo API key and use client.responses.create. See OpenAI SDK compatibility.

What orchestration models are supported?

List engines and supported models with GET /v1/engines. Each engine entry includes per-model token pricing.

Can AI assistants generate videos?

Yes. Pipevideo provides an MCP server with tools for listing engines, generating videos, checking status, and viewing usage.

How do I add credits?

Sign in to pipevideo.co, open Billing, and top up your organization balance.

Where is my data stored?

Generated videos are stored securely and served via signed URLs. See Data handling for retention and privacy details.

How do I get help?

See Help and support for contact options and what to include in a support request.