Subprocessors
Last updated August 25, 2026
A subprocessor is a third-party service we use to help run Ducktate. Because transcription and your notes stay on your own device, the list is short. This page names every service we rely on, what it does, and the data involved, so you can see exactly who touches what. It also lists the optional AI providers you can connect under your own account.
Services we use to run Ducktate
These are engaged by Gravitas AI LLC. Some run only when you use a specific feature (for example, a payment provider runs only when you buy). All are located in the United States.
| Provider | What it does | Data involved |
|---|---|---|
| Supabase | Accounts, authentication, subscription/entitlement records, and storage for problem reports you send | Account email, authentication identifiers, entitlement status; feedback text and any diagnostic bundle you choose to send |
| Stripe | Payment processing for purchases made on our website | Payment and billing details you enter at checkout (we never see your full card number) |
| Apple | Payment processing for purchases made in the iOS app | Purchase and transaction data handled by Apple under your Apple Account |
| RevenueCat | Manages App Store subscriptions | Product, transaction, and entitlement information tied to an anonymous identifier |
| PostHog | Product analytics for the app and website (on by default; you can turn the app's analytics off in Settings) | Anonymous usage events, pageviews, and crash/error diagnostics; never your dictation text, transcripts, or notes |
| Cloudflare | Hosts the website and serves app downloads | Standard request data (such as IP address) needed to deliver the site and files |
| Resend | Sends transactional and invite email | Your email address and the contents of the message we send you |
| Slack | Notifies our team internally when you send feedback or a problem report | The feedback or report you chose to send |
| Hugging Face | Hosts optional higher-accuracy transcription models you can download | None of your content: only a one-time download of model files |
AI providers you connect yourself
AI cleanup is optional, and you choose the engine. When you connect one of these, you do so under your own account or API key, and your text is handled by that provider under your agreement with them, not ours. We never send your content to these providers on your behalf, and we never see or store your credentials for them.
| Provider | When it runs | Data involved |
|---|---|---|
| OpenAI (ChatGPT) | Only if you select ChatGPT and sign in with your own account | The transcript text and limited note context, sent under your account |
| Anthropic (Claude) | Only if you select Claude, using the Claude command-line tool installed on your Mac | The transcript text, processed under your own Anthropic account |
| OpenRouter | Only if you are on Pro and connect your own OpenRouter API key | The transcript text and limited note context, sent under your key |
| Your own sync server | Only if you turn on vault sync and enter your own server address and key | Your note content, sent to a server you operate (not to us) |
Running a local model instead keeps AI cleanup entirely on your Mac, with no provider involved.
Changes to this list
We may add, remove, or replace subprocessors as the product evolves. When we do, we will update this page and its "Last updated" date. For how this fits our broader data practices, see our Privacy Policy.