Autoplay
Last updated: October 5, 2026
Account data: your email address and profile details, via Firebase Authentication. Content you create: scripts, prompts, reference images, generated images and videos, and project data, stored in Google Cloud (Firestore and Cloud Storage). Voice data: voice samples you upload, record or link, the voice models cloned from them, and recordings you supply to drive a render. Likeness data: photos you use to train an Identity Lock, and the likeness models trained from them. Voice recordings and face likenesses can be biometric-adjacent data, so we treat them with extra care (see sections 3 and 4). Usage data: generation counts and job status, used to operate limits and recover interrupted renders.
We use your data to run the studio: authenticate you, store your projects, fulfil the generations you request, recover paid renders after interruptions, enforce per-account limits, and provide support. We do not sell your personal data, and we do not use your content to train our own models.
Voice samples for cloning are sent only when you choose to give a character a voice or to clone a singing voice in the music studio. Character voices: samples are stored in your account's storage area and sent to our voice providers (Fish Audio, ElevenLabs) solely to create a private clone and synthesize the dialogue you direct; the clone samples stay with those voice providers. When a shot is rendered to a line, the synthesized dialogue is sent to our media generation infrastructure provider to drive that shot. These clones are marked private, scoped to your account, and never published or shared with other users. Deleting a voice model in the studio's voice manager deletes it from our voice provider; removing a voice from a character unlinks it without deleting the model, and remaining provider-side clone copies are purged on request. Recordings you supply as a line: when you upload or record audio to drive a render yourself — re-voicing a clip in the fine-tuning studio, or an audio reference in AI Video — that recording is sent, as you provided it, to our media generation infrastructure provider to drive that render. Only supply recordings of your own voice or of someone who has given you permission. Singing-voice clones (music studio): the sample you provide is sent to Suno, through its API partner, to create a singing voice for your songs. That clone is held on Suno's service under its terms — we cannot mark it private the way we do character voices. Removing it from the music studio removes it from your workspace only; provider-side copies are purged on request.
Identity Lock trains a reusable likeness of one person for the Portrait image models. Training requires you to confirm that the photos are of you or that you have the person's written permission; we record that attestation with the lock, and licensed talent must come through the Talent marketplace. The training photos are sent to our media generation infrastructure provider, which builds and keeps a likeness model scoped to your account so your renders can reuse it; it is never published or shared with other users. Deleting the lock in the studio immediately stops any render from using it and asks the provider to delete the model; any remaining provider-side copy is purged on request.
Infrastructure: Google Firebase / Google Cloud (auth, database, file storage, functions). Image and video generation: our media generation infrastructure provider, which receives the prompts, reference images, Identity Lock training photos, clips and audio needed for each render and passes them to the developer of the model you choose (for example Kling, Seedance or FLUX.2). Voice and audio: Fish Audio (voice clones and dialogue), ElevenLabs (narration voices, sound effects, and speech-to-text for captions of your recordings and for quality checks) and Suno (music and singing-voice clones, through an API partner). Writing, planning, message drafting and image understanding: Anthropic (Claude). Voice agents and messaging, if you use them: Google (Gemini realtime voice), ElevenLabs (conversational voice), Twilio (calls and SMS) and SendGrid (email). Payments (where offered): Stripe — we never see your full card details. Each processor handles data under its own privacy terms.
Your projects and media stay in your account until you delete them. Deleting a project removes it from your workspace immediately; the underlying media files and any provider-side copies are purged on request or on periodic cleanup (typically within 30 days). To request full account deletion — including all stored media, voice models and likeness models — contact us at the address below.
Access to your data is gated by Firebase Authentication; paid endpoints require a verified sign-in token; voice models are ownership-checked server-side; and media URLs served from storage are access-scoped. No system is perfectly secure — report suspected issues to the contact below and we will investigate promptly.
You can access, export, correct, or delete your data by using the studio's own controls or by emailing tyler.murray@autoorder.io. Depending on where you live, you may have additional statutory rights (GDPR, CCPA, and similar); we honor verified requests under those laws. Material changes to this policy will be reflected in the date above.