Blur Personal Info Before Screen Sharing: The Privacy Toggle
The open-context Dashboard has a one-click privacy toggle that blurs personally identifiable fields — your role, industry, background, interests, and current focus — with a CSS blur filter, so you can share your screen on a call, demo, or recording without exposing anything personal. It's a pure display change: your underlying preferences are never edited, redacted, or sent anywhere when you flip it.
Key takeaways
- The toggle lives in the Dashboard header as an eye icon button labeled "Hide info" / "Show info".
- It blurs role, industry, background, interests, current focus, and conversation titles.
- Tech profile (languages, frameworks, tools) and communication style stay visible — they're not treated as PII.
- Blurring is client-side CSS state only; nothing is written to your context store or sent over the network.
- The setting doesn't persist — it resets to off on every page reload, so you re-enable it each time you're about to share your screen.
Why your AI preferences dashboard needs a privacy switch
The whole point of the open-context Dashboard is to surface exactly the personal context Claude needs to be useful — your role, your industry, what you're working on, how you like to communicate. That's also exactly the kind of information you don't want visible the moment you open a screen share for a pair programming session, a conference demo, or a support call. Unlike a code editor, where you control what's on screen line by line, a preferences dashboard is designed to put your personal context front and center. The privacy toggle exists to resolve that tension without forcing you to log out or clear your data first.
What the privacy toggle actually blurs
Clicking "Hide info" applies a blur filter to specific fields across the Dashboard's preference cards and conversation list. A yellow banner confirms it's active: "Personal information is hidden — safe to share your screen."
| Card / section | Fields blurred |
|---|---|
| Work & Personal | Role, industry, background, interests |
| Current Focus | Top of mind, projects, goals |
| Conversations list | Conversation titles |
What stays visible on purpose
Two cards are left untouched: Tech Profile (experience level, languages, frameworks, tools) and Communication Style (tone, detail level, response format, language preference). open-context treats these as technical working preferences rather than personally identifying information, so they stay readable even in privacy mode — useful if you're screen sharing specifically to show off your tech stack setup without revealing who you are or what you're building.
How the toggle works under the hood
There's no server round trip involved. The Dashboard component keeps a single boolean,
privacyMode, in local React state. Fields that should be redactable are wrapped in a
small Redacted component that conditionally applies Tailwind's blur utility along
with select-none, so blurred text can't be copy-pasted out from under the filter
while it's active. Toggling the switch just flips that one boolean — nothing is read from or
written to your ~/.opencontext context store, and no network request fires.
When you'll actually use this
The most common moment is right before you share your screen: a live coding session, a demo of the tool itself, a support call where someone's helping you debug your setup, or a recorded walkthrough for a blog post or video. In each case you want the interface and the mechanics visible — but not your name, industry, or the specific project you're currently focused on. One click before you start sharing, one click after you stop, and you're back to normal.
What the privacy toggle doesn't do
It's worth being precise about the boundaries. The toggle doesn't encrypt anything at rest, and
it has no effect on exported markdown files, the Preferences editor, or the Conversations import
view — it's scoped to the Dashboard route only. If you need your exported preferences.md
or memory.md to omit certain details entirely, edit those fields before exporting;
the blur is a screen-share safeguard, not a data-minimization tool. For broader questions about
keeping your AI context off someone else's server in the first place, see our post on
self-hosted vs. cloud AI memory.
Import your context and try the Dashboard's privacy toggle yourself.
Get started with open-context →FAQ
What fields does the open-context privacy toggle hide?
It blurs your role, industry, personal background, interests, current focus (top of mind, projects, goals), and conversation titles on the Dashboard. Tech profile fields and communication style stay visible.
Does the privacy toggle delete or change my saved data?
No. It only applies a CSS blur to the affected text in your browser. Your preferences data on disk or in your configured context store is never edited, redacted, or transmitted anywhere when you toggle it.
Does the privacy setting stay on after I refresh the page?
No. The toggle is local component state in the Dashboard, so it resets to off every time you reload the page. Turn it on again right before you start sharing your screen.
Why are my tech profile and communication style still visible when I turn it on?
open-context treats languages, frameworks, tools, tone, and detail level as non-identifying technical preferences rather than personally identifiable information, so they are intentionally left out of the blur.
Does the privacy toggle work in the CLI or MCP server too?
No, it's a web UI-only feature built into the Dashboard component. The CLI and MCP server don't render a screen and have no equivalent blur state.