Privacy

Your data, and what Brink does with it.

Brink only looks at what it needs to judge focus during a session — and most of it never touches disk. Here is exactly what is read, what is kept, and what is thrown away.

Closed beta privacy policy

Where your data goes

Brink only looks at what it needs to keep you on task — and most of it is gone in seconds.

The text on a page you open → Checked in the moment, then thrown away within two minutes
Explicit content signals → Never saved as images — sensitive evidence is minimized or discarded
Page addresses and titles → Kept and encrypted, so you can review a session later
When you’re not in a session → Brink reads nothing at all

Encrypted, never sold, and never used to train unrelated models. Focus checks run with AI partners like Anthropic.

What Brink collects

Brink stores account details, session settings, consent records, heartbeat metadata, adjudication records, and Stripe identifiers needed to operate a focus session. During a session, the Brink app checks window titles, app identity, web addresses, and open-document text so it can decide whether your activity matches the focus contract you declared. The browser extension may read the active tab’s URL, page title, and bounded visible page text so it can judge whether a page is on-topic and apply browser blocks. Current metadata mode does not take or upload screenshots.

Browser extension data

The extension does not read pages when no Brink session is active, and it skips Brink’s own pages and trusted essential sites such as Gmail. When page text is needed, Brink keeps it in temporary server memory for up to 120 seconds for adjudication, then discards it — retaining only non-reversible hashes, character counts, URLs, titles, timestamps, and diagnostic metadata for session review and support. The extension stores its session token and connection settings in the browser profile so it can stay connected.

Sensitive content

Brink minimizes sensitive material. If a future evidence mode ever needs to inspect image-like evidence, explicit or pornographic content is never written to storage; Brink hashes and discards it in memory and logs only the hash. In the current metadata mode, Brink does not store raw images of screen content.

How your data is used

Beta data is used to run sessions, adjudicate focus, block off-topic content, support appeals, debug the monitoring system, and improve reliability. To perform that adjudication, Brink may send eligible page text, document text, URLs, titles, app/window metadata, and focus topics to AI service providers such as Anthropic. Brink does not sell your data, and does not use private session content to train unrelated models.

Retention and access

Brink may retain encrypted session records, adjudication records, and diagnostic metadata needed to run sessions, support appeals, debug reliability issues, and audit contested outcomes. Internal access is limited to operational review, support, debugging, appeals, and compliance work.