Windows 10 or later, minimum 8GB RAM, NVIDIA GPU with 4GB+ VRAM, at least 15GB of free disk space. AMD and Intel Arc GPUs are not supported yet. macOS and Linux support are planned — hardware costs money, so every donation gets us there faster.
No. NyxVox runs fully offline after installation. Network is only used when you explicitly trigger it — web searches, Telegram integration, or optional model updates. Nothing happens in the background.
Download all parts into the same folder. Run NyxVox_Setup.exe — the installer finds the rest automatically and handles everything from there.
Yes. Click More info → Run anyway. The warning appears because code signing certificates cost money — working on it. A VirusTotal report is available on request. Every donation helps get there faster.
Each file has a SHA256 hash listed on the download page. On Windows, open PowerShell and run: Get-FileHash NyxVox_Setup_Part1.exe -Algorithm SHA256 — compare the output against the listed hash.
Completely. Your conversations, documents, and data are stored in an AES-256 encrypted local database. Nothing leaves your machine. See the Privacy Policy for full details.
Ghost in the Wire™ is free, always. Full local AI, web search, Telegram, encrypted storage, one-click install. Spectre in the Machine™ is where the latest features land first — for those who want more. Ghost never gets left behind — paid features rotate down over time.
Yes — with Spectre tier. Bring your own GGUF model. Universal model support (exl, any quantization) is on the roadmap.
Spectre is in early access. Drop your request in the #feature-requests channel on Discord — that's where we're managing the queue and deciding what ships next.
The #roadmap channel on Discord is where I share what's coming, what's being considered, and my vision for NyxVox.