Privacy: the Cast this page extension
Effective . Applies to the browser extension only. How the service works
The short version
The extension sends a page to us only when you click the button on that page, to prepare a quote and turn it into an audio episode in your own private podcast feed. Work you start can finish after you close the popup. The extension does not monitor your browsing or capture pages you haven't opened it on.
What is sent, and when
- When you open the connected popup on a page: the page's text and markup, address and title are sent to our server to prepare a quote. This happens before you choose to generate audio. Executable scripts, styles and form fields are removed from the copy. The remaining markup can include content beyond the article itself.
- Pages you are signed in to: the capture can include subscriber-only or otherwise private content already visible in your browser. Only open the extension on a page you want us to process.
- When you click "Add to my feed": your approved price, and your image-narration choice.
The extension sends no browsing history, cookies from other sites or analytics events. Reopening a recent preview can reuse its quote; reopening a recording checks its progress. Closing the popup does not cancel a capture or recording request you have already started.
How it authenticates
Connecting the extension mints a scoped access token stored only in your browser's extension storage. It can price and produce episodes for your account, and nothing else: it cannot read your account, change settings, or create more tokens. You can revoke any connected browser in Settings, which cuts that browser off immediately.
What we keep
- In your browser, the extension stores recent page addresses, quote titles and prices, your image-description choices and recording progress. This lets it resume when reopened and avoid repeating an unresolved recording request. It does not store captured page HTML in extension storage. Older completed previews are cleared as you use it; unresolved recording records stay available until resolved or extension data is cleared.
- On our server, captured page content is stored so your episode can be produced, retried if production fails, and supported if something reads wrong. It is tied to your account and never appears on any public surface: not in samples, not in anyone else's feed.
- Preview sources expire and are removed within seven days. Once an episode is produced, its source is retained with the episode for support and retries.
- We keep the extracted text, speaking parts and voice assignments, and production quality records so we can investigate missing content, incorrect voices and other reported problems.
- The finished audio lives in your private, token-addressed podcast feed.
- Deleting your account (Delete my account in the app's Settings, or by email) removes your captured pages, episodes and the rest of your data within 7 days; copies in our backups roll off within a further 30 days. We keep only the payment records tax law requires, with your email address replaced by a reference that doesn't identify you, and a keyed one-way hash of your address so the free first article and other one-time offers can't be claimed twice.
What we never do
- Sell captured content or listening activity, or publish your captured pages in public samples.
- Use captured content for anything other than producing and supporting your episodes.
- Capture anything without a click.
Third parties
Cloudflare hosts the service and its source and audio storage, and processes generated audio for transcription checks.
OpenAI and Anthropic receive the text of the articles, scripts and documents you convert, to work out who is speaking and how numbers, dates and abbreviations should be read. They also receive article bylines to help choose a narrator and, when descriptions are enabled, relevant images, code and supporting context to describe them. Turning descriptions off stops images and code being sent; it does not stop the text being sent. Neither OpenAI nor Anthropic uses data sent through its API to train its models.
Fish Audio receives the text and voice selections needed to produce speech. Fish's terms allow it to use content submitted to its service to improve its models. That includes text from pages you send with the extension.
The service operator can inspect captured source to investigate a production problem. Each processor receives the material needed for its role.
Questions or removal
Write to accounts@askwhocasts.com, including if you want captured pages deleted from your account. To delete the whole account yourself, use Delete my account in the app's Settings.