Escalium's Advanced Video Editor works like a professional editor: a multi-layer timeline, lyrics synced word by word, color correction, background removal, text and audio effects, and Full HD export.

What makes it different is that everything happens on your own device, inside the browser.

Your videos are never uploaded to a server to be rendered: your processor, your graphics card and your memory do the work.

That has clear advantages. Your files stay private, there are no render queues, and on-device transcription doesn't use credits. In exchange, the experience depends on the computer you use. In this article we explain what you need for the editor to run smoothly.

Quick summary

Below we explain the reason behind each point.

1. Browser: the requirement you can't skip

The editor uses modern browser technologies. If one is missing, some features won't be available:

• WebCodecs is the export engine. It encodes video as H.264 and audio as AAC directly on your device. Without WebCodecs you can't export. Chrome, Edge and Safari 16.4 or later have it. In Safari before version 26, audio is encoded a different way (WebAssembly), so it still works, just a little slower.

• WebGL2 powers color correction, LUTs and scopes. If your browser doesn't have it, the editor will tell you it can't export a video with color corrections.

• WebGPU is needed for the most accurate transcription model (more on that below). Without it, the editor switches to the fast model automatically.

• Extra features only in Chrome and Edge:

  • Using the fonts installed on your computer in your texts.

  • Saving large exports straight to disk. From about 150 MB, the video is written to the file you choose instead of being assembled in memory. That's more stable and uses less RAM.

  • Choosing a folder for your created videos.

Our recommendation: use the latest version of Google Chrome or Microsoft Edge on a computer, or the Escalium desktop app. They give you every feature of the editor.

2. Processor (CPU)

The processor handles a lot of the work:

• Composing every frame of the export: video layers, animated text, transitions and effects, at 30 frames per second. The Advanced Editor composes frames on a single main thread, so single-core performance matters a lot, not just how many cores you have.

• Mixing the audio from every layer.

• "Fast" transcription (Whisper Base), which runs on the CPU.

• Converting ProRes and DNxHD/DNxHR files. Browsers can't play these camera formats, so the editor converts them to H.264 MP4 in the background. It's a demanding job.

• Creating proxies: lightweight 540p copies of 4K or very heavy clips, so the preview plays smoothly.

Minimum: a reasonably modern 4-core processor. You can edit without problems, but projects with many layers or effects will take longer to export.
Recommended: 6–8 cores from a recent generation. On a Mac, any Apple Silicon chip (M1 or newer) runs this editor very well.

3. Memory (RAM)

RAM is what you notice most as a project grows:

• During export, full-resolution frames are kept in memory. Each one takes about 8 MB at 1080p.

• If your browser can't write to disk while exporting (Safari, Firefox or a phone), the whole MP4 is assembled in memory before it downloads.

• ProRes/DNx conversions (up to 20 minutes of footage) also happen in memory.

• The AI models (transcription and background removal) are loaded into memory when you use them.

Minimum: 8 GB. With less, and especially on devices that report 2 GB or less, the browser may close the tab during long exports.
Recommended: 16 GB or more, especially if you work with 4K clips, ProRes footage or long projects with many layers.

Tip: close tabs and apps you aren't using before a long export.

4. Graphics card (GPU)

The graphics card plays three roles in the editor:

1. Hardware video encoding. The editor always asks for your graphics card's H.264 encoder, which is several times faster than the processor and uses very little battery. Almost every computer from the last few years has one, including Intel, AMD and Apple integrated graphics. If yours doesn't, the editor uses the processor: the export comes out the same, just slower.

2. Color correction (WebGL2). Color wheels, curves, LUTs, masks and scopes are computed on the GPU. Any modern integrated graphics is enough.

3. "Accurate" transcription (WebGPU). The Whisper Large V3 Turbo model runs on the graphics card. With a dedicated GPU, or Apple Silicon, it transcribes a song in very little time.

Background removal and text behind the subject (MediaPipe segmentation) are also faster with a good GPU. On modest graphics, the preview uses a lighter model so playback stays smooth. The export always uses the highest-quality model.

Minimum: integrated graphics with WebGL2 (Intel UHD 620 or newer, AMD Radeon Vega integrated graphics, Apple M1).
Recommended: WebGPU-capable graphics with 4 GB of VRAM or more (NVIDIA GeForce GTX 1650 / RTX 3050 or better, AMD Radeon RX 6000 or better, Intel Arc, or Apple Silicon), so you can use accurate transcription without waiting.

5. Storage

The editor keeps several things on your device, inside the space your browser gives Escalium:

• Transcription models: about 760 MB for the accurate one and about 140 MB for the fast one. They're downloaded once and kept.

• Library clip cache: up to 1 GB, so clips you've already used aren't downloaded again.

• Your own videos, proxies and thumbnails. Your clips are never uploaded: they stay on your computer.

• The videos you export: one minute at 1080p takes about 120 MB.

Minimum: 3 GB free.
Recommended: 10 GB free or more, on an SSD. With an SSD, clips load and move around the timeline much faster.

6. Screen and workspace

The editor brings the library, preview, inspector and multi-layer timeline together on one screen. On small screens the panels stack so everything fits, but editing is more comfortable with room to work.

• Minimum: 1366 × 768.

• Recommended: 1920 × 1080 or higher.

• Extra: with two monitors, you can move panels to the second one, just like in professional editors. That leaves your main screen free for the preview and the timeline.

7. What about phones and tablets?

The editor opens on phones and tablets and exports on the device itself. Even so, it's designed for computers. On a phone, the whole video is assembled in memory, so long or heavy exports can fail. Some mobile browsers (older iOS versions or Firefox for Android) can't export at all. For projects with several layers, color correction or heavy clips, use a computer.

8. What the editor exports

So you know what to expect from your computer:

• Resolutions: 720p and 1080p (plus custom sizes and per-platform presets).

• Aspect ratios: vertical 9:16 and horizontal 16:9.

• Frame rate: 30 fps.

• File: MP4 (H.264 + AAC), compatible with TikTok, Instagram, YouTube and any player.

• Source footage: 4K clips are supported. The editor creates lightweight proxies so editing stays smooth, but the export always uses the original clip at full quality.

• ProRes and DNxHD/DNxHR: converted automatically to H.264 (up to 1080p) so you can edit them.

9. Tips for a smoother experience

1. Use an up-to-date Chrome or Edge, or the Escalium desktop app.

2. Plug your laptop in before exporting. On battery, many computers limit CPU and GPU power.

3. Let the editor create proxies for your 4K or very heavy clips.

4. Export one video at a time. The editor already queues exports, because several at once just compete for the same resources.

5. Download the transcription models on a good connection the first time. After that they work without downloading again.

6. If your computer is on the limit, use the fast transcription model or cloud transcription (it uses credits), and export at 720p.

In short

If you have a computer from the last five or six years with 8 GB of RAM and an up-to-date browser, you can use the Advanced Video Editor. With 16 GB of RAM, a modern 6–8 core processor, a WebGPU-capable graphics card and an SSD, editing and exporting will be fast even on big projects, and you can use the most accurate transcription without waiting.