[REDACTED] home

Blog ·

What Is Video Redaction? How It Works

Short answer: video redaction is hiding private or sensitive information in a video before you share it. One software vendor describes it as "removing or masking identifiable information from videos", and a face, a license plate, text on a screen or a voice can each be that information.

This page gives general information, not legal advice. We read the linked pages on November 18, 2026. Our tools run in your browser, so your video stays on your device or secure network and never touches our servers.

What can show private data in a video?

A video holds a picture for every frame and, often, a sound track. Private data can be in either one.

  • Faces. The people in the picture, including bystanders who did not choose to be filmed.
  • License plates. A plate is a number that can connect a clip to one vehicle.
  • Text on a screen. A screen recording or a filmed monitor can show emails, names, API keys and passwords, sometimes for half a second.
  • Speech. Names, addresses and other details that people say aloud, and the sound of a voice.
  • Hidden file data. A video file can carry data such as the location and the date. Our Video Face Blur page says that its new video does not copy the original's location and date data.

Two public sources show that this list is not only our view. A vendor guide says that organizations "typically redact" faces of individuals, license plates, computer screens and credit card numbers, and that redaction covers "even audio". CaseGuard says that its software "automatically blurs faces, plates, screens, & more". These are vendors that sell redaction software, so read them as descriptions of the work, not as neutral research.

Why do people redact video?

People redact video to protect bystanders, to keep a secret out of a demo, or to meet a rule. We do not know which rules apply to you. Two public texts show how the law can name this kind of data:

  • The U.S. HIPAA rule on de-identification, 45 CFR 164.514, lists identifiers that a covered entity removes under one of its methods. The list includes "Full face photographic images and any comparable images", "Vehicle identifiers and serial numbers, including license plate numbers" and "Biometric identifiers, including finger and voice prints". The rule applies to health information that a covered entity holds. It does not say that every video must be redacted.
  • The EU GDPR defines personal data as "any information relating to an identified or identifiable natural person". A person who can be identified "directly or indirectly" by an identifier, such as a name or an online identifier, counts. Article 4 does not list faces or plates by name, so we do not say that it does.

If a rule applies to your footage, ask the person who knows that rule. The tools on this site do not check footage against any law.

How does video redaction work?

A video redaction tool follows four steps. The first two are the hard part.

  1. Find. The tool or a person finds the private item. A one-minute video has more than a thousand frames, and the item moves.
  2. Follow. The cover must stay on the item from frame to frame. CaseGuard describes a tracker that follows an object "until it leaves the screen", and a manual mode where you draw a box and follow it frame by frame.
  3. Cover. The tool draws a blur, large pixels or a solid shape over the item, in each frame.
  4. Export a new video. The cover is drawn into the new file, so the original pixels of that area are not in it. The original file stays as it was.

No automatic tool finds every item. A face that is small, turned away or fast can be missed for a moment, and one missed frame can show it. Watch the new video to the end before you share it.

Blur or black box: which should you use?

SAMPLE

Original

SAMPLE

Blur

Pixelate

Black box

A drawing of made-up text. It is not a frame from a video, and it does not show the exact output of a tool.

For normal privacy, such as a bystander's face in a clip that you post, Blur and Pixelate are correct. For text that must not be read, and for high-security work, use Black box. A blur or pixelate can leave shapes that a person can recognize. Short text can also be guessed: someone who knows the font can apply the same effect to likely words and compare the results. Our Text Unredactor shows this on pixelated text. A black box replaces the pixels with one color, so there is nothing to compare. For more on this choice, see Redact vs. Blur a License Plate.

One more rule for any style: share the new, flat video file. Do not share a project file in which the cover is a layer that someone can turn off.

Which of our tools does which job?

These three tools make a new MP4 file in your browser. Your video stays on your device or secure network and never touches our servers. They are free, with no account.

  • Video Face Blur finds and tracks each face and hides it with Blur, Pixelate or Black box. You can choose which people stay visible. It hides faces only.
  • Body Cam Face Blur does the same face job for bystanders in body cam clips. All faces are hidden at the start, and you click a person to show them. Its page says that it hides faces only, and that names, badges, license plates, screens and tattoos stay.
  • Screencast Redactor reads the text of a screen recording and covers emails, phone numbers, names, API keys, passwords and faces, such as a webcam in the corner. You can also drag a box over an area for the whole video. Its page says that it can miss small, faint or fast text.

What these tools do not do

  • They do not hide speech. The new video keeps its sound when "Keep audio" is on, so voices and spoken names stay. Turn "Keep audio" off to make a video with no sound. The tools do not find and remove only some words in the audio.
  • Video Face Blur hides faces only. It does not hide text, plates, screens or documents. For text on a screen recording, use the Screencast Redactor.
  • No tool in this set hides plates in a video. License Plate Blur works on still images. It does not work on video.
  • Length and size. Each video can be up to 10 minutes long. A video larger than 1080p is made smaller to 1080p. For a longer video, split it into parts of 10 minutes or less in a video editor. The Body Cam Face Blur page says that it makes one new file for each part and does not join them.
  • No automatic check of the result. Play the new video to the end, and look for anything that the tool does not cover.

Where a paid redaction suite is better

A suite such as CaseGuard or Redactor.ai covers more than our tools. CaseGuard says that its software blurs faces, plates, notepads and screens, has bulk redaction across hundreds of videos, and works with audio. The Redactor.ai guide lists faces, license plates, computer screens, credit card numbers and audio. If you must redact hundreds of videos, plates in video, or spoken words, those suites do jobs that our tools do not. We did not test them, and we do not claim that they are better or worse in quality.

CaseGuard also warns that free online tools "require you to upload your video to an external server for processing", and says that its own software is installed on your machine or network. Our tools do not upload your video either: the page has no server code, and the work happens in your browser.

How to check a redacted video before you share it

  1. Play the whole new file, not only the start.
  2. Step through the parts with many people, fast movement or dark scenes, one frame at a time in a player that can do this.
  3. Look for plates, screens, name tags and text that the camera draws in the picture.
  4. Listen for names, addresses and other details that people say.
  5. Share the new MP4 file, and keep the original in a safe place.

Last read: November 18, 2026. Vendor pages and rules change, so check the linked pages before you act.