Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.usevelo.ai/llms.txt

Use this file to discover all available pages before exploring further.

Creating a VeloTwin takes just a few minutes, but the quality of your photo and voice recording directly affects how natural your AI presenter looks and sounds. These tips will help you get it right the first time.

​
Face capture

Lighting Good lighting is the single biggest factor in photo quality. Use natural light from a window in front of you, or a ring light aimed at your face. Avoid backlighting - if the light source is behind you, your face will appear dark and the avatar will look flat. Camera position Position your camera at eye level. Looking up or down at the camera produces an unflattering angle that the avatar will replicate. If you’re using a laptop, raise it on a stand so the lens is level with your eyes. Face framing Center your face in the frame with your head taking up roughly the middle third of the image. Include your shoulders if possible - it looks more natural than a very tight crop on just your face. Expression Look directly into the camera with a neutral, natural expression - the kind you’d have in a professional headshot. Avoid smiling too broadly or looking off to the side. Velo will add natural movement like blinking and nodding automatically. Background Use a clean, uncluttered background. A plain wall or softly blurred background works well. Avoid busy patterns, other people in the frame, or bright objects directly behind your head. You’ll have the option to keep or change the background after the photo is taken.

​
Voice recording

Environment Record in the quietest space you can find. Background noise - air conditioning, traffic, other people talking - will be picked up and can reduce clone quality. A small room with soft furnishings (like a home office or bedroom) tends to work better than a large open space. Microphone Use your best available microphone. A dedicated USB or XLR mic will produce the best results, but a quality headset mic or even a recent laptop microphone can work well. Avoid using your phone’s speakerphone or a microphone that’s very far from your mouth. How to read the script Read the training script at your natural speaking pace - the pace you’d actually use when presenting. Don’t rush, but don’t slow down artificially either. Speak with your normal energy and expressiveness. A flat, monotone reading will produce a flat-sounding clone. Consistency Try to maintain a consistent volume and distance from the microphone throughout the recording. Moving toward or away from the mic mid-recording will create variations that reduce clone quality. One take is usually enough If the system accepts your recording, you’re good. Don’t over-think it. The quality assessment will tell you if there’s a problem worth fixing.