Support
We're here to help.
Email is the fastest way to reach us. We respond to every customer message within one business day - usually the same day.
Common questions
How do I install Northbeams?
Northbeams is one platform with three install paths: browser extension,
Mac app, PC app. Install one or all three from
/download: the Chrome Web Store listing
(Chrome, Edge, Brave, Arc), a signed .pkg for Mac, and a
signed .msi for PC. After install, sign into your dashboard
at monitor.northbeams.com and
each surface pairs to your workspace automatically through the same sign-in.
The desktop apps also watch CLI tools (Claude Code, Aider) running on the
same laptop. No extra install for CLI coverage.
Can I deploy the desktop apps to a fleet via MDM?
Yes. The signed .pkg (Mac) and signed .msi (PC)
are MDM-deployable on Intune, Jamf, and Kandji. The install token is
short-lived and signed, so push deployments enroll into the right workspace
without per-machine manual steps. Email
hello@northbeams.com for the
fleet enrollment guide and a sample MDM profile.
Why isn't the dashboard showing my detections?
The most common cause is that a surface hasn't paired yet. Open Settings in the dashboard and check that each install you expect to see (browser extension, Mac app, PC app) shows as Connected. If a surface is installed but not paired, sign out and back in on that surface. If everything shows Connected and you're still not seeing detections after visiting an AI tool or running a CLI agent, email hello@northbeams.com and we'll take a look.
Does Northbeams see my actual prompts?
No. The browser extension classifies prompts locally, inside your
browser via the extension's content script. The original prompt text
never leaves your device. Only category labels (e.g.,
credentials detected), per-pattern match counts, and a
redacted snippet (with detected secrets masked as
[REDACTED:type]) are transmitted.
The desktop apps don't see prompts at all. They watch outbound connection metadata (which AI hosts your laptop talks to) and process names (which AI desktop apps and CLI tools are running, e.g. Claude Desktop or Cursor). No prompt content, no keystrokes, no screen capture. See our Privacy Policy for the full per-surface data flow.
How do I disable sensitive-content detection?
In the browser: right-click the Northbeams icon in your toolbar → Options → uncheck “Detect when sensitive content is pasted into AI tools.” Tool-visit detection (the lighter-weight check) continues unless you uninstall the extension.
On Mac and PC: the desktop app never inspects prompt content, so there is nothing to disable. Open the menu-bar (Mac) or system-tray (PC) icon and quit the app to pause outbound-connection and process-name watching, or uninstall it from System Settings (Mac) / Add or Remove Programs (PC).
How do I cancel my subscription?
Go to Settings → Billing in the dashboard and click Manage billing. That opens the Stripe Customer Portal where you can cancel, change seats, or update your payment method. You'll keep access through the end of your current billing period.
What's free, and what costs money?
Every workspace starts on Sentinel (the top tier) free for 14 days, no card required. After day 14, the workspace falls back to Beam: aggregate-only AI visibility (tool inventory, severity totals, last 7 days, no per-user data, no exports). Beam is capped at 10 active users and runs for 60 days, then the workspace pauses unless you subscribe. Lighthouse is $15 per user / month (per-user attribution, blocking, exports); Sentinel is $25 per user / month (audit logs, MDM, compliance exports). Both have 20% off on annual billing. See pricing for the full breakdown.
I represent a company with specific security or compliance requirements.
Email hello@northbeams.com with your needs (SSO/SAML, BAA, on-prem classifier, custom retention, audit log export, etc.) and we'll set up a call.
Reporting a security vulnerability
Please email security@northbeams.com with details of the vulnerability and a reproduction. Do not file a public GitHub issue or post on social media. We aim to acknowledge within 48 hours and patch high-severity issues within 7 days.