Glossary
The words behind the work.
Making video without a camera comes with its own vocabulary. Here are the 30 terms you'll meet most, explained in plain language with no jargon and no fine print.
People on screen
- AI avatar
- In Kyndrify, an Avatar is a synthetic character that can present a script on screen. It does not represent a real person, so it is different from a consented Digital Twin.
- AI talking head
- A video with a presenter speaking directly to the viewer. The presenter can be a real-person Digital Twin or a synthetic Avatar, and each Render should be reviewed before it is published.
- Digital Twin
- A saved real-person identity built from reference images, a voice, and current consent. You can reuse a Digital Twin while it remains ready and consent is valid, but each output can still vary.
- AI presenter
- An on-screen presenter that delivers a script without a new camera take. In Kyndrify, it can be a consented Digital Twin or a synthetic Avatar, depending on whether it represents a real person.
- AI spokesperson
- An on-screen presenter used to represent a brand or message across multiple videos. Use a consented Digital Twin for a real person or an Avatar for a synthetic character, then review each Render for consistency.
- Digital human
- A broad term for a realistic person shown through synthetic media. Kyndrify keeps two cases separate: a Digital Twin represents a real person with consent, while an Avatar is a synthetic character.
- AI UGC ad
- A short ad that uses the casual style of a creator or customer review. The presenter can be a consented Digital Twin or a synthetic Avatar. It should not be presented as an unscripted customer testimonial, and any required AI disclosure still applies.
Voice and sound
- Voice cloning
- Creating a synthetic voice from an authorized recording so it can read supported scripts. It aims to keep recognizable qualities of the source voice, but delivery can vary and each output needs review.
- Voiceover / text-to-speech
- Turning written words into spoken audio. You can use a ready-made voice or an authorized voice you enrolled, without recording every script in a booth.
- Lip sync
- Matching a presenter's mouth movements to the spoken words. The goal is natural timing, but the match can vary and should be checked before the video is published.
- AI dubbing
- Replacing the spoken track of a video with a new voice, often in another language, while aiming to preserve the original timing. Review pronunciation, meaning, and lip sync before publishing.
- Voice library / stock voices
- A collection of ready-made voices you can choose instead of enrolling your own. Pick a voice that fits the message, then review the generated delivery before using it.
Footage and motion
- B-roll
- Extra supporting footage that plays over narration, like a product close-up or a city street. It adds visual context when the main presenter is not on screen.
- Image-to-motion
- Using a still character image and a short reference video to render a new moving clip. The image supplies the subject, while the reference video supplies the motion.
- UGC video
- A short, casual clip made in the style of someone speaking on a phone. UGC means user-generated content, so a synthetic version should not be presented as an unscripted customer testimonial.
- Product photography (AI)
- Studio-style product concepts rendered from a product reference without arranging a new photo shoot for each idea. Check labels, edges, and product details before using an output in shops or ads.
- Faceless video
- A video that tells its story with footage, captions, and a voiceover instead of a presenter on camera. It is one option when the message does not need an on-screen person.
Making and output
- Storyboard / script
- The written plan for a video: what is said and, sometimes, what is shown in each part. A clear script keeps the message and shot choices easier to review.
- Render
- Kyndrify's word for both submitting supported inputs to produce an output and the resulting file. A Render can be video, image, or audio, depending on the Studio tool.
- AI video generator
- A tool that renders video from inputs such as text, images, audio, or saved characters. Studio uses guided controls and defaults, so you can render clips without working on a full timeline.
- Text-to-video
- Rendering moving video from written instructions. Depending on the tool, the text may describe a visual scene, spoken script, or both, and the output still needs review.
- AI video editor
- Software that helps assemble or change video clips. Kyndrify currently provides guided render tools rather than a full drag-and-drop editing timeline, so final assembly may happen in another editor.
- Aspect ratio
- The shape of a video frame, such as tall for phone feeds or wide for websites and presentations. Pick one that matches where the video will be shown.
Trust and disclosure
- AI disclosure
- A clear notice that media was made or changed with AI. Platform and local rules vary. Kyndrify's built-in disclosure and provenance coverage is still expanding across output types, so check the export and add a visible label when required.
- Content Credentials (C2PA)
- An open standard for attaching signed claims about a file's source and edit history. The credential can be inspected or removed, and its presence does not prove that the content itself is true. Kyndrify is still rolling this layer out across output types.
- Synthetic media
- An image, audio track, or video that is partly or fully generated or altered by software. A Digital Twin Render and an Avatar Render can both be synthetic media even though their consent rules are different.
- Deepfake
- Media that convincingly changes or generates a person's face or voice, especially without permission or to mislead. Current consent and clear disclosure help separate legitimate synthetic media from deceptive use.
- Consent-first AI
- A design approach that requires current permission before using a real person's face or voice. In Kyndrify, this applies to real-person Digital Twins. Synthetic Avatars do not use a real person's likeness and follow separate rules.
- Likeness rights
- A person's legal rights and control over how their face, voice, or identity is used. Kyndrify requires users to confirm the needed rights before rendering with a real person's likeness.
- Invisible watermark
- A hidden signal added to media so it may remain detectable after some common edits or reposting. It is not a guarantee of traceability. Kyndrify's invisible-watermark coverage is still being rolled out and is not yet universal.
Now make your own.
You've got the vocabulary. Put it to work on the free plan: one full minute on us, no credit card required.