AI Research

Research behind the talking head.

Clone13's product is built on a generative-video pipeline that brings photo animation, voice cloning, and lip-sync into one workflow. The generative models are licensed from specialist providers; what we build and tune is the workflow around them. Here is what we work on and why it matters for automotive retail.

Effective November 6, 2026

Photo animation

Turning a single static headshot into a natural, on-camera talking head — keeping identity, expression and head movement as faithful as the model allows, and catching the cases where it does not.

Voice cloning

Voice cloning from a short sample, aiming to keep the speaker's accent, tone and cadence across languages. Results differ by sample quality and by language.

Lip-sync

Aligning generated speech to mouth movement so the picture follows the words instead of drifting behind them.

Multilingual synthesis

A single cloned voice speaking in more than one language while keeping as much of the speaker's timbre as the model can hold.

Consent & provenance

We are exploring watermarking and provenance signals so generated material can be traced, and we keep a timestamped consent record for every likeness in the product today.

Evaluation

We review output for realism, identity fidelity and lip-sync, and we tune the experience from what dealerships report. We do not publish benchmark figures we cannot stand behind.

Our principles

  • • Consent first. Every photo and voice sample is uploaded by its owner after signing our consent, scoped to their dealership, and destroyed on the schedule in our Biometric Retention & Destruction Policy.
  • • Narrow deployment. We build for dealership sales communication — not open-ended impersonation.
  • • Transparency. A Clone13 video is a salesperson’s introduction, shared by their own dealership — not a synthetic statement from someone else.
  • • Honest claims. We describe what the pipeline does today and where it still needs work, rather than promising a perfect replica of a camera.

Get in touch

We’re a Chicago-based product team. If you’re a researcher or partner interested in responsible generative video for automotive retail, write to hello@clone13.com.