LOCKE FOR PERPLEXITY
Ask Perplexity real questions without putting real data in them.
Perplexity reframes AI as search, and people query it the way they query Google — fast, conversational, and often loaded with specifics. "What does this lab result mean for someone with my history," "is this clause in my lease enforceable," "how do I appeal a charge on account 4012..." The research framing invites users to include exactly the personal details they'd never type into a public search bar, because it feels like a private assistant.
But the query doesn't stay private. It's processed by upstream models and used to drive retrieval against external sources to compose an answer. The sensitive particulars that make your question specific are the same particulars that leave your device — folded into a request whose whole purpose is to reach outward.
Locke lets you keep the specificity without the exposure. It detects the personal and confidential details inside your query and masks them on-device, so Perplexity still answers the question you actually have — just without carrying your private facts out onto the open web.
The risk with Perplexity
Perplexity's search framing encourages people to put highly personal specifics into their questions — health, legal, and financial details — and that query is then processed by models and used to retrieve from the open web, carrying those details off the device.
How Locke helps
Locke detects and masks the sensitive details inside a query before it's submitted, so Perplexity answers the substance of your question without your private facts leaving the device. Locke — the Sonomos desktop app, coming soon — extends this across every AI tool you use, not just one.
Keep using Perplexity — without the exposure
Locke runs entirely on your device. Sensitive data is detected and masked before any prompt is sent, so nothing confidential ever leaves your machine. Pricing for Locke, the desktop app, is coming soon; Canary is free and open source.
