Architecture overview

Frontend (React 19, SPA)

Single-page React app, served as static files from Railway's edge. The editor is roughly 30k lines of JavaScript organised into ~250 components. State is managed by React hooks + Context (no Redux). Routing via react-router-dom.

All video work happens inside the browser via three native APIs:

• WebCodecs — H.264/H.265/VP9 encode + decode, hardware-accelerated when the GPU supports it. Used for export on Pro tier.
• MediaRecorder — fallback export when WebCodecs isn't available (older browsers, Safari pre-17).
• FFmpeg.wasm — handles unsupported input formats (MOV, AVI, WMV, MKV, FLV) by transcoding them to a browser-playable codec. Runs in a Web Worker spawned by the library, off the main thread.

Does AetherCut process videos on its servers?

No. AetherCut has no server side video processing pipeline and no video processing endpoint. The backend handles accounts, payments, analytics, email, and opt in AI proxy calls without retaining audio or prompts beyond the call.

The backend's only jobs are:

• User accounts (Google OAuth + session cookies)
• Stripe payments + webhook handling
• Anonymous analytics (page-views collection)
• Proxy for opt-in AI features (Whisper, Claude, Kling, Nano Banana) — your audio/prompts pass through; we don't retain them beyond the call
• Email transactional flows via Resend (welcome, retention drips you can unsubscribe from)

The backend has no video processing pipeline. There is no POST /api/video/process endpoint. If we wanted to, we couldn't process your video without shipping new code first.

Does AetherCut store my video in the cloud?

No. AetherCut has no cloud video storage bucket, no server side FFmpeg, no CDN video transformations, and no third party video processing APIs. It also does not use session replay tools that record your screen.

• No cloud video storage (S3, R2, etc.) — we have no bucket your footage could live in.
• No server-side FFmpeg — only the browser-side .wasm copy.
• No CDN-side video transformations.
• No third-party video processing APIs.
• No analytics SaaS (Mixpanel, Amplitude, Segment, Heap) — we use our own minimal page-view collection so the data never leaves our database.
• No session-replay tools (FullStory, LogRocket, Hotjar). These tools record your screen; we do not.

Can I inspect AetherCut's editor code in production?

Yes. AetherCut ships bundle source maps with every build, so you can open DevTools, inspect the main JavaScript bundle, and read the editor code with variable names and function bodies intact.

Bundle source maps ship with every build. Open DevTools → Sources → look at main.[hash].js. You can read the actual editor code — variable names, function bodies, the lot. Most apps strip source maps in production. We don't.

What actually leaves your device during normal use

In a typical editing session, almost none of the heavy data moves over the network. When you drop a video into the editor, the file is read by the browser from local disk, decoded locally, and either played directly or converted in-browser first if the format is not natively usable. Timeline edits, trims, cuts, captions, and export all happen against memory and browser APIs on your machine rather than against a remote render queue. The practical implication is simple: your internet speed does not determine how fast a video export finishes, and disconnecting from the internet does not suddenly stop the core editor from working once the app is loaded.

What does still cross the network is the ordinary account and product plumbing around the editor. Signing in creates a session, upgrading hits Stripe, loading the app itself fetches static assets, and opting into AI features sends the specific audio or prompt needed for that call through the backend proxy. That distinction matters because people often assume any slowdown is “the server processing my footage.” In this architecture, a slow export is usually local: unsupported browser, weak hardware acceleration, limited RAM, thermal throttling, or a fallback path such as MediaRecorder or FFmpeg.wasm doing more work on your device.

A good way to verify this yourself is to open your browser’s Network panel before importing a large clip. You should not see a giant upload request when you start editing or exporting in the normal non-AI workflow. You may see small API calls for auth, billing, or page loads, but not a multi-hundred-megabyte transfer of source footage to an editor backend. That is the clearest operational test of the architecture: if there is no upload endpoint receiving your video, there is nowhere server-side processing could be happening.