Privacy

Last updated: October 8, 2026.

Cutscene is designed so that your source code never reaches us. This page explains what we do receive, where it goes, and how long we keep it.

What we receive

• Your answers to the questionnaire: a pitch, features, colors, fonts, copy, layout descriptions, a call to action, and the brief you gave your agent.
• Files you approve: usually a logo and a few screenshots.
• What we produce: the plan, scene code, music and the rendered video.
• Job records: timestamps, job state, and usage numbers for each render.

Your coding agent reads your repository on your machine and decides what to send, following the Cutscene skill's rules: no source files, config, secrets, internal URLs or customer data. You approve the full package before anything is sent. Review it — your agent, not Cutscene, has access to your code.

Where it goes

• Job records are stored in Google Cloud Firestore and files in Google Cloud Storage, in the United States (us-east1).
• Videos are rendered in an isolated Daytona sandbox that is reset between jobs.
• A Claude model from Anthropic plans and builds the video, so it receives your answers and files for that job.
• Video download links are signed and expire after 7 days.

We do not sell your data, use it for advertising, or share it beyond the providers above.

How long we keep it

During the beta, job records, uploaded files and rendered videos are kept so you can download and revise your videos. Ask through the channel you received your invite from and we will delete a job and its files.

This website

This site sets no cookies and runs no analytics. Our hosting providers (Vercel and Google Cloud) keep standard request logs, such as IP addresses and timestamps, for security and operations.

Your access token

Your beta token identifies you to the Cutscene MCP server. Keep it in an environment variable, don't commit it, and ask for a new one if it leaks.