A recorder that captures what you meant
Every other screen recorder gives you pixels. DemoRiff records the pixels and the intent behind them — which element you clicked, what you typed, when a modal opened, where a page navigated. That structured record is why the edit can be automatic instead of approximate.
- 4K60
- Maximum capture
- 4 tracks
- Screen, camera, mic, system
- < 3%
- CPU overhead while recording
- 0 setup
- Runs in the browser
Interaction data, not guesswork
Clicks are bound to DOM elements and accessibility labels. The pipeline knows you pressed “New shipment”, so the zoom frames the button and the guide writes the step correctly.
Separate tracks, always
Screen, camera, microphone and system audio are recorded independently, so you can drop the camera, mute the app, or replace your voice later without re-recording.
Redaction at capture time
Sensitive fields can be masked as they are recorded, not after. What never enters the file cannot leak from it.
Record in a tab, a window, or the whole desktop
The browser extension records any tab at up to 4K without installing anything on a locked-down machine. The desktop apps for macOS and Windows add full-desktop capture, hardware encoding, system audio on Windows, and offline recording that uploads when you reconnect. Both produce identical downstream results.
- Chrome and Edge extension — no install privileges required
- macOS and Windows desktop apps with hardware encoding
- Offline recording with resumable background upload
- Pause, resume, and re-take a single step without starting over
The metadata that makes automatic editing possible
Alongside the video, the recorder writes a structured event log: element clicked with its selector and accessible name, text typed (redacted by policy), scroll position, route change, modal open and close, network idle. That log is what lets the pipeline frame a zoom on the right element, split the recording into logical steps, and write a guide that says “Click New shipment” rather than “Click at 1420, 84”.
- Element selector, role and accessible name per interaction
- Route and view changes as chapter boundaries
- Typed input captured as intent, with values redacted by default
- Exportable as JSON for your own tooling
Intent tracking requires the browser extension or desktop app. Imported video falls back to cursor-motion inference, which is good but not exact.
Blur it before it is ever written to disk
Point at a field and mark it always-masked, or write a rule — any element matching a selector, any text matching a pattern, any region of the frame. Built-in detectors already cover emails, phone numbers, card numbers, API keys, JWTs and common customer-name fields. On Enterprise, masking policies are set centrally and cannot be disabled by the person recording.
- Element, region, regex and detector-based masking
- Masked at capture time, so raw values never persist
- Workspace-wide policies that individuals cannot override
- Audit record of every masking rule applied to every recording
A presence that does not fight the product
Camera is recorded as its own track with background removal, so you can place it as a circle, a rounded rectangle, a full-frame intro, or nothing at all — and change your mind after the fact. Eye-line correction keeps you looking at the viewer while you read the script off a second monitor.
- Background removal without a green screen
- Eye-contact correction, including with glasses
- Auto-hide when the cursor enters the camera region
- Replaceable with an AI presenter later, same track
The technical detail
- Resolution
- Up to 3840×2160 at 60fps (4K on paid plans)
- Capture targets
- Browser tab, application window, full desktop, multi-display
- Audio
- Microphone and system audio as separate 48kHz tracks
- Platforms
- Chrome, Edge, macOS 13+, Windows 11
- Max length
- 4 hours continuous, resumable
- Storage
- Encrypted at rest, region-pinned on Scale and Enterprise
Our solutions engineers were spending a third of their week rebuilding the same demo for different verticals. Now they record once and the variants generate themselves. We gave the team back twelve hours each.
About screen studio recorder
No. The browser extension covers most product demos and needs no admin rights. The desktop apps exist for full-desktop capture, system audio on Windows, and recording native applications, but everything downstream works identically either way.
That is what the masking policies are for. Most teams record against staging with a seeded demo tenant, but when production is unavoidable, workspace-level masking rules apply automatically and cannot be turned off by the person recording. Values are masked before the frame is encoded, so they are never written to storage.
Pause, say “let me take that again”, and continue. The pipeline detects retakes and keeps the later attempt. You can also mark a step for re-recording afterwards and splice in a fresh capture of just that step — everything downstream rebuilds.
Pairs well with
In production at