Cluely is easily detectable / detector source code

Honrly, a local-first, self-hostable interview-integrity tool, has released its source code, demonstrating that AI cheating tools like Cluely and Interview Coder are easily detectable. The tool scans running processes in real time and flags known cheating apps, aiming to ensure fair interviews by rewarding preparation and honest effort.

A local-first, self-hostable interview-integrity tool built for a simple reality: "undetectable" tools are not magic. Apps like Cluely , LockedIN , Interview Coder , and Parkaeet market themselves as invisible, but that is marketing fiction, not a reality, THEY ARE SUPER EASY TO DETECT. FEEL FREE TO FIND IT OUT for yourself : Honrly runs on your own machine and does two things: Self-Monitor — scans your running processes in real time and flags known AI cheating tools such as Cluely and Interview Coder , plus any custom apps you add to a watchlist. Everything happens locally; nothing is uploaded. Video Call Preview — shows exactly how a monitored video call will look, then lets you connect a real room from a hosted provider Daily.co or your own self-hosted service e.g. Jitsi via iframe . There are no accounts, no login, and no company database . This is a clean, hackable starting point you can build on. Interviews should reward preparation, skill, and honest effort. People put real time into studying, practicing, and showing up ready; they deserve a fair shot without being undercut by tools that try to turn cheating into a product. Tauri 2 Rust desktop shell React 18 + Vite + TypeScript frontend Tailwind CSS + a small set of Radix UI primitives @daily-co/daily-js for the optional hosted video provider The Rust backend exposes two local commands src-tauri/src/main.rs : list processes — enumerates running processes ps on macOS/Linux, tasklist on Windows . get input activity — reports recent keyboard/mouse activity macOS implemented . The frontend src/screens/LocalMonitor.tsx polls list processes every few seconds and applies simple, fully local heuristics to flag suspicious apps. Add your own rules in isSuspiciousProcess , or add app names at runtime via the Custom Apps panel. Open Video Call Preview and either: Daily.co — paste a room URL like https://your-team.daily.co/room-name . Self-hosted — paste any iframe-embeddable meeting URL e.g. a Jitsi room on your own domain . Without a URL you still get a full layout preview with mock participants and call controls, so you can see the experience before wiring up a backend. Tip: run the Self-Monitorin one window while a call is live to demonstrate live cheating-tool detection. install JS deps npm install run the web frontend only browser, no native process scanning npm run dev run the full desktop app requires the Rust toolchain npm run tauri:dev build a production desktop bundle npm run tauri:build Process scanning relies on the Tauri Rust backend, so use npm run tauri:dev to exercise the Self-Monitor. In a plain browser npm run dev the monitor will show a permission/availability error because the native list processes command isn't present. - Node.js 18+ - Rust toolchain rustup and the Tauri prerequisites https://tauri.app/start/prerequisites/ for your OS src/ React + Vite frontend App.tsx routes: / Home , /monitor, /video screens/Home.tsx landing page screens/LocalMonitor.tsx process scan + Cluely detection screens/VideoCallScreen.tsx call preview + connect Daily/self-hosted components/ui/ small Tailwind/Radix UI primitives + VideoCall src-tauri/ Tauri Rust backend src/main.rs list processes, get input activity GNU Affero General Public License v3.0 AGPL-3.0 . See LICENSE /niravrohra/Honrly-app/blob/main/LICENSE .