Research interviews, news footage, workplace investigations and everyday social posts often feature people who should not be identifiable. A face is the obvious giveaway, but a lanyard, a car, a laptop screen or the street outside a window can point to someone just as surely.
This page walks through each of those, and the tool above covers them on your device so the raw footage never goes to a third party. It changes the picture only, so read the section on sound before you share anything.
Keep the original somewhere secure and open the video in the tool above. Your original file is left as it is; the anonymized version is saved as a new file.
Choose Black box, Pixelate or Blur, then Chosen areas and Find faces automatically. Every face it detects gets a moving oval that shows while that person is in view.
Drag a box over each badge, plate, screen or sign, track it if it moves, and time it with Starts here and Ends here. For a revealing background, use the switch that applies the effect to everything except the areas.
Watch the whole clip at full screen, stepping frame by frame where people move quickly. Then select Export video, and play the exported file once more before you share it.
Before you start, list everything in the footage that could point to a person. Typical items include:
Faces are the only item found automatically. You have to spot and cover everything else on the list yourself, which is why a slow, careful review matters more than any setting.
Find faces automatically scans the clip with an on-device detector and joins each person's detections into a single moving oval. Look for anyone it skipped, such as a person in profile or far away, and draw an oval over each one. Someone who faces away for the whole clip may still be known by their hair or build, so judge whether they need covering too.
A badge moves with the wearer and turns as they do. Box it with some margin, select Track movement automatically, and scrub through the moments when the person turns or folds their arms.
Give each plate and house number its own area, shown only while it is readable. The license plate guide explains following a plate through dashcam footage.
Screens tend to hold the most exact identifying details, so cover them with Black box. The guide to hiding text shows how to time a box to a notification or a scrolling chat.
For anything that travels across the frame, the guide to blurring a moving object explains what the tracker can follow and how to correct it.
Sometimes the place gives a person away more than their face does: a recognizable office, a classroom wall display, the view from a window. When the subject must stay visible but the surroundings must not, draw an area around the subject and turn on the switch for everything except the areas. Everything outside that shape is then covered.
When nobody in the clip should be identifiable, choose Whole video instead. A strong blur or large blocks over every frame keeps the movement and the general scene while removing the detail. The background blurring guide covers the keep-one-area mode in more depth.
For anonymity, the question is not which effect looks best but which leaves the least behind. Black box replaces the area with solid black, so there is nothing to sharpen, guess or rebuild. Choose it when the people involved could face harm or retaliation if recognized.
Pixelation with big blocks hides a face well while still showing viewers that someone is there. The tool shows a warning below 12 px, and anything that identifies a person deserves far larger blocks than that.
Blur reads as the gentlest option and is common for incidental passers-by in social posts. It is also the easiest to under-do, so keep the strength high and judge it at full size.
Each export uses a single effect for all of its areas. To combine black boxes on badges with blurred faces, run the clip through twice, adding the second set of areas to the first export.
Knowing the limits matters as much as using the tool well.
None of this is legal advice. Privacy and data protection rules differ between countries and between uses, so ask whoever handles data protection in your organization if you are unsure.
Open the video above, pick Black box or a strong Pixelate, and select Chosen areas. Use Find faces automatically, add boxes for badges, plates and screens, then review the clip and select Export video.
No. The audio is kept as recorded, apart from a format conversion for WebM. You can remove the soundtrack completely under Export settings, but there is no voice changer or pitch shift.
Sometimes, but not always. A light blur can leave a face recognizable to people who know it, and clothing, location or voice can give a person away on their own. Black box plus a careful check of every other detail is the safest approach.
Covering identifying details can reduce the personal data a video holds. Whether footage really counts as anonymous depends on everything else that could identify people, including the sound and the context, so take advice from your data protection contact for anything formal.
Find faces automatically handles many faces in one pass, each with its own oval. Distant or partly hidden faces are easy to miss, though, so for a dense crowd consider Whole video with a strong effect.
No. The video stays in your browser, which matters for research data, source material and HR evidence that should not reach outside services.
Not from a black box, because those pixels are gone from the new file. Weak blur or small blocks leave more behind, and your untouched original still exists, so keep it stored securely.