Subprocessors
To run Enqlave we rely on a small number of trusted service providers (“subprocessors”) that process data on our behalf. We keep this list short by design, and we publish it so you always know who is involved.
| Subprocessor | Purpose | What they process | Location |
|---|---|---|---|
| Supabase, Inc. | Account sign-in, encrypted library sync and storage, diagnostics | Account identifiers; your library only as encrypted ciphertext we cannot read; diagnostic counts, durations, and error codes | United States |
| Google LLC (Gemini API) | Reads and describes documents you import, builds the index behind search by meaning, and answers questions — when you have configured a Gemini key | Document content, extracts used for the search index, and your questions — sent directly from your device, using your own API key, at the moment of each call | United States |
| Anthropic, PBC (Claude API) | Reads and describes documents, and answers questions — when you have configured a Claude key | Document content and your questions, sent directly from your device, using your own API key, at the moment of each call | United States |
| OpenRouter, Inc. | Carries the same AI work — reading, indexing and answering — for people who have not configured a key of their own | Document content, extracts used for the search index, and your questions — sent directly from your device, at the moment of each call, under a key we issue for your account. OpenRouter is a router: it passes each request to a model provider, which performs the work | United States |
| Apple Inc. | App distribution (TestFlight) and Sign in with Apple | Your Apple ID identifier and basic install/device information | United States |
Enqlave sends document content to an AI provider at three moments: when a document is first read and described, when extracts from it are turned into the vectors behind search by meaning, and when you ask a question. The first two happen on their own, as part of filing a document — you don’t have to ask for them. In every case the request goes straight from your device to the provider, not through a server we operate, and we keep no copy of what is sent.
Which provider receives it depends on your keys. If you have configured your own Gemini or Claude key, that key is used and the request goes to that provider — this is unchanged. If you have configured no key, Enqlave uses a key we issue and pay for, and the request goes to OpenRouter, which passes it to a model provider. Requests on that path carry a zero-data-retention flag, which restricts them to endpoints that do not retain what they are sent.
Your stored library is encrypted on your device before it syncs; our cloud holds it as ciphertext it cannot read. That holds on both paths — our servers issue the key, they never see the document.
Our marketing website is hosted by Netlify, Inc. and the domain is registered with Porkbun, LLC. These handle only website data — never your documents.
We’ll update this page before adding or changing a subprocessor. Questions: hello@enclave-ai.com.
← Back to home