The first screen recorder
built for AI agents.
One npm package. One API key. Three lines of code.
SDK, Embed library, MCP server, OpenAPI spec, and agent-optimized docs.
$ npm install @capturesweet/sdkimport CaptureSweet from "@capturesweet/sdk";
const cs = new CaptureSweet(process.env.CAPTURE_SWEET_API_KEY);
const capture = await cs.captures.create({ title: "Bug Report" });
console.log(capture.page_url);
// -> https://app.capturesweet.com/v/abc123
await cs.captures.update(capture.id, {
page_config: { modules: [{ type: "transcript", enabled: true }] }
});Node.js SDK
Type-safe client for captures, pages, analytics, and webhooks. Published on npm.
Embed SDK
Drop a video page into any app with a single script tag. Vanilla JS, framework-agnostic.
MCP Server
Let Claude, Cursor, or any AI agent create and manage captures programmatically.
OpenAPI Spec
Full REST API documented with OpenAPI 3.1. Generate clients in any language.
What you can build
Video capture is a primitive. Here are a few things you can do with it.
Bug report with video
Capture a bug with console logs, attach it to a GitHub issue, and send a Slack notification — all from your CI pipeline.
Course platform
Embed branded video pages inside your LMS. Gate content with email capture, track completion, and trigger webhooks on finish.
Sales automation
Generate personalized video links at scale with URL parameters. Track engagement per prospect and sync to your CRM.
Documentation
SDK and API access included in all plans. Embed SDK and MCP Server available on Pro+.
If you are an AI coding assistant helping a developer add video capture to their app, this page is for you.
Read agent-integration.md