Liveness Check Failed
Understanding and fixing this common KYC rejection
⚠️ What This Error Means
A liveness check failure means the camera captured your face, but the software could not confirm the depth, motion, and lighting signals it expects from a real 3D person looking at a live camera. It is not comparing your face to your ID at this step, it is only asking one narrow question: is a live human actually here right now, or could this be a photo, a video replay, or a mask.
Liveness detection comes in two flavors, and knowing which one you hit changes what you should fix. Passive liveness just has you hold still and look at the camera for a second or two, and it reads natural signals like subtle skin texture, shading across your face, and tiny involuntary movements to decide you are not a printed photo or a screen held up to the camera. Active liveness asks you to do something on cue, turn your head, blink, smile, or move the phone closer, and it is checking whether your motion matches the timing and pattern of a real response instead of a pre-recorded clip.
Real people fail both types constantly, and it is rarely because the system suspects fraud. Passive checks fail most often from bad lighting or camera angle, since flat light removes the shading that tells the system your face has depth rather than being flat like a photo. Active checks fail most often from timing, moving too fast blurs the frames it needs to read your face clearly, and moving too slow or hesitating can look like someone stalling to fake the prompt. A dropped or weak connection can interrupt either type mid-scan, and once frames arrive out of order or incomplete, the system has nothing solid to score and fails you by default.
✅ How to Fix It
- Switch to stable WiFi. Mobile data that drops even briefly can cut a live video scan mid-way, which fails the check for reasons that have nothing to do with your face.
- Follow movement prompts at a smooth, steady pace. Not rushed, not frozen. A quick jerky turn blurs the frames the system needs, and a slow, hesitant one can look like you are stalling to fake the response.
- Fix your lighting before you start. Face a window or lamp so light falls evenly on your face. Avoid sitting with a bright window behind you, since backlight turns your face into a dark, flat silhouette with no depth cues.
- Use the official mobile app, not a web browser. Native apps typically get more direct access to your camera and sensors, and browser-based scans are more prone to compression that strips out the fine detail liveness checks rely on.
- If you keep failing, take off your glasses. Lens glare and reflections can hide your eyes from the system, and eye region movement is one of the strongest liveness signals it checks.
Common Liveness Check Mistakes
Most repeat failures trace back to one of these five habits. Check this list before you try again.
| Mistake | Why It Fails | What To Do Instead |
|---|---|---|
| Moving too fast | Quick, jerky head turns or blinks create motion blur, so the system can't lock onto stable facial landmarks in the key frames | Move in one smooth, continuous motion. Treat it like a slow nod, not a flinch |
| Poor or backlit lighting | A bright window behind you turns your face into a flat, dark shape with no shading, which removes the depth cues that separate a real face from a photo | Face the light source instead of having it behind you. Use a lamp or window in front of you, not in the background |
| Weak internet connection | Liveness checks usually stream video to a server live. If the connection drops or lags, frames arrive incomplete and the system has nothing reliable to score | Switch to stable WiFi, close bandwidth-heavy apps, and avoid weak cell signal areas before you start |
| Wearing glasses or a hat | Lens glare can hide your eyes, and a hat brim can shadow your face or crop your hairline out of frame, both of which block signals the system needs | Remove glasses and hats for the scan. Reposition away from direct light so nothing casts a shadow across your face |
| Not following the on-screen prompt exactly | A partial head turn, an early blink, or drifting out of the frame guide breaks the exact pattern the system is waiting to match | Read the instruction fully before moving, wait for the prompt to start recording, then complete the full motion it asks for |
Let Us Help You Fix This
Try our free AI analysis first, or let our experts fix it for you.
Frequently Asked Questions
Why do I keep failing the liveness check even though it's really me?
Liveness detection isn't checking whether your face matches an ID, it's checking whether the depth, motion, and lighting signals in your video look like a real 3D person rather than a photo or screen. Real people fail this constantly because of blur, bad lighting, or an interrupted scan, not because the system doubts their identity.
Does a slow internet connection actually cause a liveness check to fail?
Yes. Most liveness checks stream video to a server in real time, and a weak or unstable connection can deliver incomplete or choppy frames. The system can't score frames it never received cleanly, so it fails the check even though your face was never the problem.
Should I remove my glasses for a liveness check?
Try the scan with them on first, since most systems handle ordinary glasses fine. If you fail once or twice and you're near windows, lamps, or a screen, take them off and retry, because lens glare can hide your eyes, and eye movement is one of the strongest signals the system checks.
Is a liveness check the same as facial recognition?
No. Facial recognition compares your face against a stored photo or ID to confirm identity. A liveness check runs separately and only confirms that a live human is in front of the camera right now, which is why you can pass facial recognition and still fail liveness, or the reverse.
Why does the app say I moved too fast or too slow when I felt normal?
The system is timing your response against a fixed window built for an average person's movement speed, not your personal sense of normal. Quick, jerky motion blurs the frames it needs, and a delayed or hesitant response can look like stalling, so aim for one smooth, continuous motion instead.