Privacy Policy

• Account information: If you sign in with Google, we receive your basic account identifiers (such as your email address and account ID) to identify your account. Your subscription status and your text token balance are stored with your account.
• Local settings and queue: Your prompts, settings and reference images are stored locally in your browser, so your work persists between sessions. When a prompt uses a reference image, the extension attaches it inside your own Google Flow project, just as you would by hand.
• Generated media: Images and videos you generate in Google Flow are downloaded directly to your own computer. In the Audio to Video studio, copies are also kept in your browser's local storage so the studio can assemble your video. For the preview in the Queue tab, a small copy (about 512 pixels wide) of each downloaded result is kept in your browser's local storage until you start a new run or clear the queue. We never receive or store your generated media.
• Audio to Video: audio and subtitles. The voiceover and subtitle files you add are stored and processed only in your browser. Speech recognition runs on your own device. Your audio is never uploaded to us or to anyone else.
• Audio to Video: text sent to our servers. When you choose to prepare scenes, the extension sends text to our servers: the transcript of your voiceover with its timestamps, the scene texts, the style and direction you chose, and the names and notes you wrote for your reference images. The reference images themselves are not sent.
• Payments: When you upgrade or buy text tokens, you complete the payment on Stripe's checkout page. We pass your account ID and email address to Stripe so the payment is linked to your account. We never see or store your card details.

Added in version 1.5

• debugger new
  Google Flow does not reliably start a generation from an ordinary automated click. The extension uses this permission for one thing only: to move the mouse to the Generate button and click it in your own Google Flow project tab, once per prompt, while a run you started is active. It attaches only to a Google Flow project tab and detaches right after each click. It does not read network traffic, cookies, page content or any other tab. While it is attached, Chrome shows a bar saying the extension started debugging the browser. This is expected and cannot be hidden.
• Optional access to flow-content.google, *.googleusercontent.com and storage.googleapis.com new
  These are the addresses where Google Flow serves your generated images and videos. The extension asks for this access only when you start generating in the Audio to Video studio, or when you press "Show previews" in the Queue tab. It uses it to fetch your own results, to attach each one to the right scene in the studio, or to show a small preview of each downloaded result in the Queue. You can decline; downloads and the rest of the extension keep working.

Core permissions

• Access to labs.google and flow.google.com, scripting, activeTab: to type your prompts into Google Flow, find the Generate button, and see when a generation has started or finished, on the Flow tab you are working in.
• storage and unlimitedStorage: to keep your queue, settings and Audio to Video projects in your browser. Unlimited storage is needed because one project can hold many images and a long voiceover.
• downloads: to save each result to your computer with ordered, numbered file names.
• identity: to sign you in with Google through your ViralDNA account.
• sidePanel: the extension's interface is shown in the browser side panel.
• notifications: to tell you when a long run finishes or needs your attention.
• alarms: to keep a long run on schedule so the queue does not stall.
• power: to keep your computer awake during a long unattended run.

• Google Sign-In (OAuth): used only if you choose to sign in, to verify your identity. Governed by Google's own privacy policy.
• Authentication backend (Supabase): securely manages your account, sign-in session, subscription status and token balance.
• ViralDNA servers (viraldna.io): check your subscription and token balance, and receive the Audio to Video text described in section 1.
• Google Gemini API: our servers send the Audio to Video text to Google's Gemini API to proofread the transcript and write the scene prompts. Only text is sent, never your audio or images. Governed by Google's terms.
• Hugging Face: the first time you transcribe, the speech recognition model (Whisper) is downloaded from huggingface.co and then kept in your browser. This is a plain file download: no audio, text or account information is sent. As with any download, Hugging Face can see your IP address.
• Stripe: processes subscription and token payments. Governed by Stripe's own privacy policy.