Frequently asked questions about Vyra

An always-on multimodal desktop assistant: voice interaction, computer vision, smart-home control, 3D CAD generation, autonomous web automation and personal memory — unified in one desktop app.

Two layers. Episodic memory records events and makes them searchable by keyword and by meaning. The world model keeps structured knowledge about your people, projects and domains.

A consolidation pass runs like a sleep cycle — important things stay sharp, noise decays.

You state a goal; Vyra breaks it into key results and tasks. A background executor works on them every 15 minutes while you're idle, and you get morning and evening briefings on progress.

Real-time voice runs on Gemini native audio. Heavy thinking routes to large reasoning models, quick tasks to fast ones — a 5-tier router picks between Local, Fast, Creative, Thinking and Ultra per request.

A local model tier (via Ollama) provides offline fallback, so core assistance keeps working without internet. Cloud models take over when connectivity returns.

Memory lives on your machine — local databases and JSON, not a cloud store. Face authentication runs fully offline as a local biometric gate.

n8n workflows end to end, smart lights and plugs, autonomous browsing with live screenshot preview, structured web scraping, and messaging on WhatsApp, Telegram and Discord.

Yes — hand tracking maps pinch, grab and fist gestures to window moves, clicks and drags. Minority Report mode, on your desk.

Describe a part in plain language and Vyra generates a parametric 3D model, previews it in-app, then slices and sends it straight to your 3D printer — printers are auto-discovered on your LAN.

Vyra scans context every 60 seconds — system load, battery, idle time, time of day — and fires helpful nudges through 7 proactive rules with a cooldown, so it helps without spamming.