The offcuff.ai Capability Matrix

Engineered to Win Every High-Stakes Call

Every feature in offcuff.ai is purpose-built to outperform Parakeet, LockedIn, and Cluely. From hardware-level display masking to sub-100ms multi-LLM reasoning.

Hardware-Level Invisibility

Gaze-Aligned Stealth Teleprompter HUD

100% Invisible to Zoom, Teams & Screen Shares

Our desktop overlay hooks directly into the Windows Desktop Window Manager (WDA_EXCLUDEFROMCAPTURE) and macOS Quartz Compositor (CGWindowSharingType.none). The overlay sits transparently directly beneath your web camera, aligning your gaze naturally with the interviewer without eye darting.

Hardware-level OS display exclusion guarantees 0% visibility on all screen-sharing platforms
Non-activating window (WS_EX_NOACTIVATE) eliminates browser blur triggers (HireVue, CoderPad)
Dynamic speech cadence word illumination highlights your pace as you speak
Customizable opacity (0% to 100%) and hotkey toggle (Cmd/Ctrl + Shift + H)
Technical SpecificationsVERIFIED 2026
Screen Capture MaskingOS Kernel Level
Browser Focus Blur0 Events Triggered
Eye Gaze Offset< 2 Degrees
HUD Hotkey ToggleInstantaneous (<5ms)
Included on all plans. No predatory paywalls or separate upgrades required.

Zero Eye-Darting Gaze Lock

Overlay positions seamlessly right under the physical webcam aperture, allowing natural eye contact while reading solutions.

Sub-100ms Inference Pipeline

Custom C++ audio capture bridge pairs with Whisper Large v3 Turbo and Claude 3.5 Sonnet to outspeed competitors by 3x.

Zero Cloud Audio Retention

Audio frames exist purely in volatile RAM ring buffers. Zero raw voice recordings are ever written to disk or third parties.

Ready to Experience the World's Most Advanced Copilot?

Launch our instant web HUD simulation or download the native desktop app with OS display masking.