--- title: "AI characters: dialogue videos, cost shown upfront" description: "Dialogue videos with fictional AI characters, a voice per character and lip sync, with the estimated cost before spending and authorisation per video." url: https://claqon.com/en/features/ai-characters/ language: en updated: 2026-10-12 alternates: es: https://claqon.com/funciones/personajes-ia/ pt: https://claqon.com/pt/funcionalidades/personagens-ia/ --- BetaAI characters # Fictional AI characters in dialogue, with the cost in view before you spend With AI characters, CLAQON produces videos in which fictional characters talk to each other: one voice per character, lip sync only in the shots that warrant it and camera movement at no cost for the rest. Before spending, you see the estimated cost and the maximum, and nothing is paid without your authorisation. It is in beta, off by default and untested with real material. [Pre-register](https://claqon.com/en/pre-registration/?modulo=funcion-personajes#configurador) [See spend control](https://claqon.com/en/features/spend-control/) - Fictional characters, never real people - Estimated and maximum cost before spending - AI disclosure when publishing ![AI Characters tab: activation switch, checklist of requirements (keys, verified prices, style, cast, cost cap and test) and the required generator notice.](https://claqon.com/recursos/img/capturas/en/personajes-1600.webp) Real CLAQON interface with sample data. What it solves ## Characters that talk without overpaying A dialogue video with animated characters needs generated images and, for the lips to follow the voice, video that is charged by the second. Without control, a mistake is expensive. With AI characters you only pay for what is needed: one image per framing, which is reused, lip sync for part of the length and FFmpeg movement for the rest. And it shows you the cost beforehand, with your authorisation for each video. Beta, not verified with real material The module is integrated and off on every account. The lip-sync rates are pending live verification, and an unverified price does not spend anything. How it works ## How a video with AI characters is prepared and produced 1. **Connect your keys** A fal.ai connection for lip sync and a Black Forest Labs connection for images, encrypted on your machine. ElevenLabs is optional. 2. **Check the price** “Check” queries the live rate of the lip-sync model. Without a check from the last 7 days, nothing is spent. 3. **Choose style and cast** A style from the catalogue or your own, and 2 or 3 characters in dialogue (up to 6 in the cast), each with their own voice and reference sheet. 4. **Set level, cap and acceptance** You choose how much of the video gets lip sync and the cap per video, and you accept the estimated cost when you switch it on. 5. **Test a shot** One frame and a few seconds of lip sync to check the vertical format, the mouth and the identity before producing. 6. **Authorise and produce** With a valid authorisation, the voice is generated first and the plan is adjusted to what has already been reserved: if it does not fit, it removes lip sync; it never asks for more. Capabilities ## Styles, lip-sync levels and spending confirmation - **17 styles**: 16 presets – 3D character series, talking objects, 2D anime, claymation or period oil painting, among others – and a custom one. Realistic ones have to be explicitly allowed. - **Custom style with filters**: It rejects studios, franchises, living artists, brands and real people, and proposes a substitute. It also blocks minors, sex and violence. - **Lip-sync levels**: Image only, Saver (lip sync for 35% of the length), Balanced (50%) and Quality (60%, with the Pro model). Lines shorter than 1.5 seconds never get lip sync. - **E and M costs**: E is the estimated cost per video and M the maximum that is reserved, according to the current price catalogue. Without a catalogue, “cost not estimable”. They are estimates, not invoices. - **Authorisation per video**: 1 to 30 videos, with a maximum for each and an expiry of 24 hours to 7 days. If a video fails without having spent anything, the authorisation is returned. - **Standing cap**: The alternative to authorising: a cap per video and the “Produce without asking me” tick box. - **AI presenter Off by default**: A lip-synced character in the hook, the re-hooks, the ending and the call to action of your narrative videos. - **Manual production**: Your own script with “Name (emotion): text” lines, with a quote before reserving. - **“Part N” series Off by default**: 2 to 10 parts; each part uses up its own authorisation. - **ElevenLabs voice per character Off by default**: Voices from your ElevenLabs account whose right of use you confirm; otherwise, local voices. - **Reused frames**: Up to 100 saved frames per account to reuse across videos. Transparency ## AI disclosure and platform labels Every AI characters video is published as AI-generated content. On YouTube, the synthetic content disclosure is sent automatically. Automatic production adds an AI footer to descriptions, except on X, and #AI on Instagram. The Instagram and Facebook API has no such field: you tick the label yourself in the app. The on-screen “AI” badge is optional and does not replace the disclosure. Each platform also applies its own labels and detection, so a video without the platforms’ AI labels cannot be promised. Limits ## Limits worth knowing about - It is in beta, off on every account and untested with real material. - Never real living people or private individuals. A historical figure only appears as an anonymous archetype, from behind or in silhouette, with no lip sync or voice. - The lip-sync rates are pending live verification, and the price catalogue expires every month: unless it is renewed, nothing paid is quoted. - Automatic production needs its workflow generator deployed; without it there is manual production, tests and the AI presenter. - There is no voice cloning or dubbing. - YouTube keeps videos uploaded through the API private until your Google project passes its audit. Frequently asked questions ## Questions about AI characters ### Can I create a character based on a real person? No. The characters are fictional: a living person or private individual never appears, and a historical figure only appears as an anonymous archetype, with no close-up, no lip sync and no voice. The custom style rejects names of real people, studios, franchises and brands. ### How much does a video with AI characters cost? It depends on the lip-sync level, the length, the style and your providers’ current rates, so there is no fixed figure. The panel calculates E and M with the current price catalogue before you authorise anything. ### Can it spend without me knowing? Not without your authorisation. You accept the cost when you switch it on and authorise each video, or you set a standing cap. A price that has not been verified in the last 7 days does not spend, and if the plan does not fit within what was reserved, it removes lip sync instead of asking for more. ### Can the videos go out without the AI label? That cannot be promised. CLAQON discloses synthetic content where the API allows it and the on-screen badge is optional, but each platform applies its own labels. On Instagram and Facebook, you tick the label yourself in the app. [Guide to labelling AI content](https://claqon.com/en/resources/labeling-ai-content/). ### What is the AI presenter? A character from your cast who appears with lip sync at the key moments of your narrative videos, such as explainers, tutorials or stories. The rest of the video keeps the account’s visual style. Early access ## Explore AI characters with spending under control AI characters is a product line in beta. If you are interested, tell us your format and we will explain what is ready and what is still to be verified. [Pre-register](https://claqon.com/en/pre-registration/?modulo=funcion-personajes#configurador) [See all features](https://claqon.com/en/features/)