I want to build a small standalone device that behaves like a live strategy-guide assistant. The idea is simple: load any game’s walkthrough onto the device, point its camera at the player’s screen, and let the device speak and display the next step only after it sees that the current step has been completed. Core requirements • Accept guides in common digital formats (PDF, TXT, HTML) pulled from the internet, SD card, or local storage. • Work with any game, regardless of platform, as long as a guide is available. • Detect on-screen progress through real-time computer-vision (camera aimed at the monitor/TV) or an equivalent capture method. • Decide where the player is in the guide and prompt the next instruction using both voice (TTS) and on-screen text. • Wait until it verifies the player has completed the action before moving on. • Small, self-contained hardware footprint—Raspberry Pi-class or similar SBC is fine as long as performance is smooth. Deliverables 1. Functional prototype: hardware schematic + fully working software image ready to flash. 2. Source code with clear build/run instructions. 3. Short demo video showing the device guiding a live play session. 4. Brief documentation outlining how to add new guides and train/adjust the vision component for different games. I’m especially interested in freelancers who can point to prior work with computer vision, natural-language processing, or embedded voice interfaces. Please highlight that experience in your reply.