Trigger any flow you've built in the visual editor from your own code. Async runs, 26 node types, Node + Python SDKs, schema-driven inputs an AI assistant can read end-to-end.
What you can build
If you can describe the pipeline as a flow, you can run it on a schedule, from a webhook, from your own app, or from a script your AI assistant wrote five minutes ago.
Loop a flow over a CMS table or RSS feed and ship a fresh short every hour without touching the editor.
Generate per-user ad variants, onboarding clips, or sales follow-ups from a single template flow.
Turn breaking stories into vertical recaps the moment they publish — script, voice, captions, render.
Spin out fifty talking-head variants in an afternoon, then push the winners to TikTok and YouTube.
Drop podcast or stream URLs in, get auto-clipped, auto-captioned shorts back as ready-to-post MP4s.
Wire video generation into your own product — webhooks, queue control, idempotent runs, cancellation.
How it works
Build the flow once in the visual editor. Drag in script, voice, captions, music, video — whatever shape your output needs.
POST /runs with inputs that match the flow's schema. The API validates eagerly and returns a run ID.
Poll GET /runs/:id or use the SDK's wait() helper. When the run completes, you get a hosted video URL.
import { Hyperclip } from "hyperclip";
const hc = new Hyperclip({
apiKey: process.env.HYPERCLIP_API_KEY,
baseUrl: process.env.HYPERCLIP_BASE_URL,
});
const { id } = await hc.runs.create({
flow_id: "<your-flow-id>",
inputs: { prompts: { 0: "a lighthouse at dawn" } },
});
const run = await hc.runs.wait(id);
console.log(run.video_url);Node catalog
Every node available in the editor runs server-side over the API. Mix and match into pipelines as deep as you need.
Build it with AI
Every flow exposes its schema. Every error has a structured code. The API is built for LLMs to introspect, validate, and self-correct — so Claude Code, Codex, Cursor and the rest can write a working integration in one prompt.
Point Claude Code, Cursor, Windsurf, or any MCP-aware client at one URL and your AI assistant gets live, searchable access to every doc page. It introspects the API instead of guessing it.
{
"mcpServers": {
"hyperclip-docs": {
"url": "https://docs.hyperclip.co/mcp"
}
}
}One click copies the entire docs site (~150–200k chars) to your clipboard. Paste into any frontier-model chat and tell it to ship a Hyperclip integration. Works with Claude Code, Codex, Cursor, Copilot — anything.
Get the bundlePaste a single doc page — quickstart, inputs reference, errors — and ask for exactly what you need. Best for targeted edits inside an existing project.
What to pasteWorks with Claude Code · Codex · Cursor · Copilot · Windsurf
Client libraries
npm install hyperclip
import { Hyperclip } from "hyperclip";
const hc = new Hyperclip({
apiKey: process.env.HYPERCLIP_API_KEY,
baseUrl: process.env.HYPERCLIP_BASE_URL,
});
const run = await hc.runs.wait(
(await hc.runs.create({ flow_id, inputs })).id
);pip install hyperclipco
from hyperclipco import Hyperclip
hc = Hyperclip(
api_key=os.environ["HYPERCLIP_API_KEY"],
base_url=os.environ["HYPERCLIP_BASE_URL"],
)
ref = hc.runs.create(flow_id=flow_id, inputs=inputs)
run = hc.runs.wait(ref["id"])Grab a key, paste the quickstart into your AI assistant, and have a working integration before the kettle boils.