Blur faces in video, automatically
Upload a video, choose to blur faces, and download a copy where every face is covered on every frame. Detection plus tracking keeps the blur stable, and the pixels are re-encoded so the original face can never be recovered.
Blurring faces in a video means destroying the pixels that make a person identifiable on every frame, then writing a new video file — not dragging a black box over a thumbnail. Done properly, the blur stays locked on each face even as people move, turn or are partly hidden, and the result cannot be reversed.
This page explains how to blur faces in a video the right way: why automatic detection beats manual editing, how tracking keeps the blur stable, and how to make the output irreversible and compliant. You can blur faces in a video right now without an account.
How automatic face blurring works
A reliable face blur is two steps working together:
- Detection finds faces in individual frames.
- Geometric tracking follows each face between frames, interpolating its position so the blur stays on target even when the detector momentarily misses it.
Detection alone is brittle: the moment someone turns their head, walks behind an object, or the frame motion-blurs, a per-frame detector loses the face and the blur flickers off — leaking the identity for a few frames. Tracking closes those gaps. Once both passes agree on where each face is across time, the pixels in those regions are blurred or pixelated and re-encoded into a new file.
Because the regions are re-encoded rather than masked, there is no hidden original layer to peel back. That is what makes the result irreversible.
Why automatic beats blurring faces manually
Manually blurring faces in a video editor is slow and error-prone:
- A one-minute clip at 30 fps is 1,800 frames. Keyframing a blur across all of them by hand is hours of work per minute of footage.
- Humans miss faces in the background, in reflections, or in fast motion — exactly the frames that leak identity.
- Many "blur" effects in editors are reversible overlays, not destructive edits, so the original is still recoverable in the project file.
Automatic detection and tracking covers every frame consistently, catches background and partially occluded faces, and writes a genuinely anonymized file. You go from a multi-hour manual task to a few minutes of processing.
Common situations where you need to blur faces in video
- Dashcam and bodycam footage — share or publish driving or incident clips without exposing bystanders, drivers or pedestrians. Pair face blur with license plate blurring so the whole scene is covered.
- Journalism and documentary — protect sources, witnesses and bystanders before publishing footage, while keeping the rest of the scene intact for context.
- Social media and user-generated content — post clips filmed in public without exposing the faces of people who never consented to appear.
- Research, training and CCTV — anonymize recordings before they are reused for analytics, model training, demos or evidence sharing.
In every case the goal is the same: keep the footage useful while removing the faces that identify people.
Don't forget the audio
Faces are only part of what identifies someone in a video. A clip can name a person, read out a phone number or state an address in the audio track — all personal data. A complete anonymization beeps or mutes those segments alongside the visual blur. Medianonymizer can blur faces and treat the audio in the same job, so you do not anonymize the picture and leak the sound.
Blur faces in your video now
Upload a video, choose to blur faces (and plates, screens or audio if you need them), see the exact price, and download an anonymized copy. No account, no subscription, pay only per job.
Frequently asked questions
- Can a blurred face be reversed or un-blurred?
- No, not when the blur is done by re-encoding the pixels rather than overlaying a mask. Medianonymizer destroys the high-frequency detail in each face region and writes a new video file, so there is no original layer underneath to recover. An overlay or a separate mask track can be peeled back and is not real anonymization — we do not use that approach.
- Is blurring faces in a video enough to comply with GDPR?
- Blurring faces irreversibly removes a direct identifier, which is a core part of anonymizing video. To take footage fully out of GDPR scope you also need to handle other identifying data — spoken names and numbers in the audio, license plates, visible documents or screens. Medianonymizer can blur faces and plates and beep or mute sensitive audio in the same job, so the whole file is treated, not just the faces.
- Which video formats and resolutions are supported?
- Common formats such as MP4, MOV, WebM and MKV are supported, and the result is re-encoded to a standard MP4. There is no hard resolution limit for typical footage; longer or higher-resolution videos simply take more processing time and are priced per minute.
- Can I blur faces in many videos at once?
- Each upload is processed as its own job, so you can run several videos in parallel by starting them one after another. There is no account and no queue to manage — you pay per job, so batching is just a matter of submitting each file. For high-volume or programmatic needs, contact us about API access.
- How much does it cost to blur faces in a video?
- Video is priced per job from €3.00 plus €0.05 per minute, and you see the exact price before you pay. There is no subscription and no account — you pay only for the videos you actually anonymize.