Mosaic Post Nothing is uploaded
Face detection and blurring both run inside your own device. Your video never travels over the network.
USE CASE
Automatically blur bystander faces in your vlogs and street videos
Filming in public means people will end up in your shots. Just pick your video and the AI detects every face and blurs it, handing you back a clip that's ready to post. Everything runs inside the browser, so your footage never goes anywhere.
Blur faces for free → No sign-up / free / your video is never uploaded
See the results for yourself
Results from the actual pipeline (left: original / right: processed).
Demo footage: Tears of Steel — (CC) Blender Foundation | mango.blender.org
Sound familiar?
- Mosaic tracking in your editor is one face at a time, and a busy street can mean dozens of them.
- You dread getting called out for a face in the frame and having to swap the video after posting.
- You don't want to repeat the same blurring chore after every shoot.
- Scenes from a trip can't be reshot — you'd hate to scrap them over a bystander in the frame.
Mosaic Post uses AI to automatically detect faces in your video and blur them, tracking each one frame by frame. It runs inside your own browser, so there is nothing to upload. However many people are in shot, and however crowded the scene, it is free — for videos up to 3 minutes.
Detection covers even the small faces deep in the frame. By using information from the surrounding frames, it captures the very moment a face appears, preventing the early-frame misses that often happen with manual tracking.
You can do the same thing in an editor, but on a street-walk video the tracking work multiplies by the number of passersby times the number of cuts, and the track slips every time someone crosses the frame or gets clipped at the edge. With automatic processing, a crowd scene with dozens of passersby takes nothing more than picking the file.
Your video never leaves your device. Detection and rendering both run in the browser, so we never receive your footage and it is never stored anywhere. There is simply no retention period to worry about. You can also go through the detected faces and choose which ones to leave visible.
Manual tracking in an editor vs Mosaic Post (automatic AI)
| Manual tracking in an editor | Mosaic Post (automatic AI) | |
|---|---|---|
| A busy street scene | Set up tracking for every single face | Every face handled automatically, same price |
| Faces entering or clipped at the edge | Tracks slip and need fixing by hand | Covered automatically using surrounding frames |
| Working time | Grows with video length × headcount | Just pick the file and wait |
| After every shoot | Repeat the same chore again | Just pick the file again (no sign-up, free) |
| Where the footage goes | Stays on your PC | Stays on your device (never sent to a server) |
How it differs from free editing apps
- The work is the same no matter how many people appear — it doesn't grow with the number of faces
- Profile views and small faces in the background are detected too. Tracked automatically across every frame, with no manual mask tweaking
- No watermark on the result, no sign-up, nothing to unlock
- Your video is never sent anywhere, so there are no deletion policies to compare
Designed to protect your privacy
Nothing to delete, because nothing is stored
We never receive your video in the first place. There is no retention window to keep track of and no deletion to wait for.
Never used for AI training
Since your video never reaches us, there is nothing to train a model on or pass to a third party. It is ruled out by the design, not by a promise.
No sign-up, no payment
It is free, and you never enter an email address or card details. Open the page, pick a video, done.
How it works
- 01
Pick a video
No account, no app to install. Open the page and choose a video — even at this point, nothing is uploaded.
- 02
Faces are found
AI detects faces across every frame, tracking everyone automatically, including people who walk in mid-scene.
- 03
Choose who to hide
Every detected face appears in a list. Take out any you want left visible — keep yourself and hide everyone around you, for instance.
- 04
Export and save
The blurred video is written out. Expect roughly 1.3 times the length of the video on a computer, and several times that on a phone.
Price
Free
For videos up to 3 minutes. Larger files, such as 4K, are scaled down to Full HD on export.
Automatic detection is never perfect. Please check the finished video before you publish it.
Frequently asked questions
Is it free no matter how many people are in the video?
Yes. Pricing is based only on the length and resolution of the video, with no surcharge for the number of people.
Can it detect faces that appear small in the frame?
Yes, even small faces deep in the frame are covered. That said, automatic detection is never perfect — please check the finished video before you publish.
Does it work on night streets and crowds?
Yes. Sample results for night-time crowd scenes are shown on our home page. Faces are detected and blurred automatically even in these conditions. Harder footage means more chance of a miss, so do check the export before publishing.
What about profile views, or people facing away?
Profile views are covered by detection. However, processing targets faces, so someone fully turned away with no face visible won't be blurred — note that the back of a head or a person's build are outside the scope.
Can I use it for videos going to platforms other than YouTube?
Yes. You receive the finished work as a video file, so it works for YouTube, Instagram, TikTok, or anywhere else — unlike YouTube Studio's blur tool, which only applies to videos uploaded to YouTube.
Will the image or audio quality degrade?
Audio is copied as-is without re-encoding. Video is exported at an ample bitrate, so there's no noticeable degradation.
Related guides
Other use cases
No sign-up, no cost. Your video is processed right here, never uploaded.
Blur faces for free →