Replace low-quality regions of your video with a sharper photo
maskOnly composites video and photos together in either direction — insert a sharp photo into matching regions of your video, or play a video inside a masked region of a photo. Automatically aligned, softly feathered, and exported as a clean MP4. No green screen, no reshoot.
See it in action
How it works
1. Upload a video and a matching photo
Start with your footage and a higher-quality still photo of the same scene — the same sky, wall, screen, or backdrop, just sharper.
2. Automatic alignment
maskOnly matches features between the photo and video frame and computes a homography to align them, correcting for crop, scale, rotation, and perspective. A manual 4-point fallback is there if a shot needs it.
3. Paint a mask, feather the edge, export
Brush, lasso, polygon, or rectangle — select the regions to replace, dial in feather and expand, and export a clean H.264 MP4.
Built for clean composites
Soft, natural edges
Per-mask feather and expand/contract controls blend the photo into the video instead of leaving a visible cutout.
Export up to 4K
Masks are stored as vector paths, not bitmaps, so the same mask renders correctly at any resolution — export up to full 4K (3840×2160).
Multiple masks, timed independently
Each mask has its own shape, feather, opacity, and start/end time, so you can composite several regions across a clip.
No reshoot required
Fix a blurry sky, screen, sign, or backdrop using a single reference photo instead of re-filming the scene.
Frequently asked questions
- What does maskOnly actually do?
- You upload a video and a higher-quality still photo of the same scene. maskOnly automatically aligns the photo to the video, then lets you paint one or more mask regions — the video plays as normal everywhere else, but inside your masks it shows the photo's sharper detail instead.
- How is this different from a green screen or chroma key?
- Chroma key replaces a solid color with different footage. maskOnly replaces a region of your existing video with high-resolution detail from a matching photo of the same scene — no green screen or second camera angle required, and the edges are feathered for a soft, natural blend instead of a hard cutout.
- How does the automatic alignment work?
- maskOnly uses feature matching (SIFT/AKAZE) and RANSAC homography estimation to map the still photo onto the video frame, correcting for differences in crop, scale, rotation, and perspective. If automatic alignment doesn't land cleanly, you can fall back to manual 4-point alignment.
- What are the feather and expand controls?
- Feather softens a mask's edge with a gaussian blur so the composite blends naturally instead of showing a hard cutout line. Expand/contract grows or shrinks the mask boundary before feathering, so you can fine-tune exactly how much of the photo bleeds into the video.
- What video and image formats are supported?
- Video: MP4, MOV, WebM. Image: JPEG, PNG, WebP (PNG alpha masks and grayscale masks are also supported for advanced masking). Output is exported as H.264 MP4.
- Are there limits on video length or resolution?
- Yes, for V1: videos up to 20 seconds, 4K (3840×2160) resolution, and 60fps; still images up to 8000px on the longest side; uploads up to 150MB.
- Do masks work at any output resolution?
- Yes. Masks are stored as resolution-independent vector paths (normalized 0–1 coordinates), not fixed-size bitmap images, so the same mask definition renders correctly whether you're previewing at 720p or exporting at full 4K resolution.
- How does pricing work?
- maskOnly uses credits — one credit renders one export. New accounts start with free credits to try it out, and additional credits are available in packs from the billing page.
- Is my footage stored or shared?
- Your uploaded video, photo, and rendered exports are stored only for your account to access your projects, and are never shared publicly.