The Biggest Update Yet
FaceStabilizer 2.0 is here, and it changes everything about how you turn horizontal footage into vertical-ready content. This is not an incremental patch or a minor tune-up. It is the largest update we have ever shipped, packed with features that our users have been requesting since day one. From an entirely new Podcast Mode to 4K export, AI-powered upscaling, and a smarter tracking engine, version 2.0 redefines what a desktop reframing tool can do.
Everything still runs 100% locally on your machine. No cloud uploads, no waiting for a server, no subscription to a rendering farm. Your footage stays on your hardware, and processing happens in real time using your own GPU. Whether you are on macOS or Windows, 2.0 is ready to go the moment you hit download.
We overhauled the entire app: new tracking, new export engine, and entirely new workflows that did not exist in 1.x. If you have been using FaceStabilizer for quick social cuts, you are going to love what is possible now. And if you are discovering the app for the first time, there has never been a better moment to jump in.
Podcast Mode: A Whole New Workflow
Podcast Mode is the headline feature of 2.0, and it deserves the spotlight. It is a dedicated workflow built from the ground up for podcast and interview footage. Import your recording, trim it down to the segment you want (up to 3 minutes per clip), and let the auto-analysis engine do the rest. FaceStabilizer detects every speaker in the frame, identifies when each person is talking, and maps the entire conversation onto a multi-lane timeline.
The timeline editor gives you full control over the output. Press S to split a segment, right-click any clip to rename a speaker or assign a unique color for easy visual tracking. Each speaker gets their own lane, so you can see exactly who is on screen and when. It is the kind of editing precision that used to require a full NLE, now condensed into a single-purpose tool that does one thing brilliantly.
When you are happy with your cuts, export individual reels per speaker. Each speaker gets their own vertical clip, perfectly framed and tracked. And with caption burn-in powered by word-level subtitles, your reels are ready to post the moment they finish rendering. No separate captioning tool, no manual syncing. The words appear on screen exactly when they are spoken, styled and timed automatically.
4K Export and Near-Lossless Quality
Version 1.x maxed out at 1080p. That was fine for most social platforms, but creators posting to YouTube Shorts, high-DPI feeds, or archiving master files needed more headroom. FaceStabilizer 2.0 Pro now supports export resolutions up to full 4K. That is 4K in every orientation, whether you are exporting 9:16, 4:5, or 1:1.
Resolution is only half the story. We also overhauled the export engine. Pro users now get near-lossless encoding, which preserves virtually all of the original detail in your footage. Skin tones stay natural, text stays sharp, and fine textures like hair and fabric do not dissolve into compression artifacts. If you have ever exported a clip and noticed it looked softer than the original, that problem is gone.
And with the new batch export feature, you can render all three aspect ratios (9:16, 4:5, and 1:1) in a single pass. Hit export once, walk away, and come back to three perfectly framed files ready for Instagram, TikTok, and LinkedIn. The Compare button lets you preview all ratios side by side before you commit, so you can verify the framing looks right everywhere.
Smart Re-Acquisition and Face Lock
Face tracking in 1.x worked well when the subject stayed in frame, but real-world footage is unpredictable. People turn away, duck below the frame, or step out of shot entirely. In version 2.0, the app handles all of this gracefully. When a tracked face leaves the frame, it holds its last known position and re-locks the moment the subject reappears. No manual intervention, no broken tracking, no jarring jumps in the output.
We upgraded the face detection engine for dramatically better accuracy in challenging conditions: low light, unusual angles, hats, glasses, masks. Tracking stays locked where other tools would lose the subject entirely.
For scenes with multiple people, the new Face Lock feature lets you pin your tracking to a specific person. Click the face you want to follow, and FaceStabilizer ignores everyone else in the frame. It is essential for conference panels, group interviews, and any multi-person setup where you need the camera to stay on one subject without drifting to whoever is closest to center.
Three Stabilization Styles
Different footage demands different tracking behavior, and a single stabilization curve cannot serve every use case. FaceStabilizer 2.0 introduces three distinct stabilization styles that let you dial in exactly the feel you want:
- Smooth — Gentle, cinematic camera movement. The virtual frame glides slowly to follow the subject, creating a polished, professional look. Best for talking-head videos, interviews, and any content where calm framing matters.
- Balanced — The default for most footage. Tracks the subject with natural responsiveness while filtering out micro-jitters and sudden jolts. A great all-around choice when you are not sure which style to pick.
- Responsive — Tight, snappy tracking that stays locked on the subject with minimal delay. Ideal for fast-moving content like sports highlights, dance videos, and action-heavy clips where the subject moves quickly across the frame.
The best part: you can switch between styles after tracking without re-analyzing the video. The tracking data is computed once and stored, so toggling between Smooth, Balanced, and Responsive is instant. Preview each style, compare the results, and export whichever feels right. No waiting, no re-processing.
AI Upscaling
Cropping into a horizontal video to create a vertical frame means you are working with fewer pixels. If your source is 1080p, a 9:16 crop might leave you with a frame that feels soft, especially on high-resolution phone screens. FaceStabilizer 2.0 solves this with built-in AI upscaling.
Two upscaling modes are available. The standard mode doubles your output resolution with impressive detail recovery, sharpening edges and restoring texture that would otherwise be lost to the crop. The High-Quality mode is tuned for footage where color accuracy and natural grain are priorities, delivering a cleaner upscale that preserves the character of the original recording.
Like everything else in FaceStabilizer, upscaling runs entirely on your local hardware. There is no upload step, no per-minute fee, and no quality cap imposed by a server. You get full-resolution, AI-enhanced output as fast as your GPU can render it.
Try It Free
FaceStabilizer 2.0 is available now on macOS and Windows. The free tier includes full face tracking, all three stabilization styles, 1080p export, and single-ratio output so you can experience the core workflow without spending a dollar. No credit card required, no account creation, and absolutely no cloud uploads.
Upgrading to Pro unlocks everything: Podcast Mode with per-speaker export and caption burn-in, full 4K resolution, batch export across all aspect ratios, AI upscaling, near-lossless encoding, and the Compare button for side-by-side ratio previews. It is a one-time purchase with no subscription, so you pay once and own it forever.
If you have been waiting for desktop reframing to truly feel like a professional tool, this is the update that delivers. Download FaceStabilizer 2.0 and see what your footage can become.