Your gateway to local AI
QVAC Workbench converts local AI into a high-quality experience that works and evolves on any of your personal devices. All the value you’re used to seeing in other AI assistants, but 100% private, truly yours. QVAC Workbench is the first real-world demonstration that personal local infinite AI is physically achievable.
Download for Desktop
Download for Mobile
A Scalable, Sovereign AI Experience That Fits In Your Pocket
Chat and experiment with local AI
QVAC Workbench brings powerful AI models directly to your device. Experiment, customize, and explore AI capabilities with full privacy and control.
No limits to your network
QVAC can delegate processing to more powerful devices you control, enabling advanced AI systems to run without compromising your privacy.
Download for Desktop
Download for Mobile
FAQ
No account creation is required. QVAC Workbench operates entirely offline with no centralized servers. All processing and data storage happens directly on your devices, ensuring complete privacy and eliminating the need for user accounts.
Nope.
Yes, Workbench supports delegated inference which allows you to offload inference to a more powerful device like a desktop or a laptop. Set it up from the app settings.
Yes, cross-device sync can be enabled from the app settings and works entirely peer-to-peer.
QVAC Workbench runs on Android, iOS, Windows, macOS, and Linux. Since all processing happens locally on your device rather than in the cloud, performance varies significantly based on your device's hardware capabilities and the models you choose to run.
This depends almost entirely on the models you are planning to run. Keep in mind that a flagship mobile with 8GB of (v)RAM or more can comfortably run models between 1B and 3B parameters.
TPS stands for Tokens Per Second and TTFT stands for Time To First Token. These are metrics that indicate how many tokens can be processed per second and how long it takes to get the first token out of the model respectively.
QVAC Workbench requires microphone permissions for voice transcription features, file access permissions to read documents you want to analyze, and storage permissions to save conversations and downloaded models locally on your device.
Local AI performance varies significantly across different devices due to hardware differences. While we test on various platforms, we can't cover every device configuration. If you experience issues, please submit a crash report or contact support with detailed information about your device and the problem you're encountering.
You can get help through multiple channels: Discord, X, and Featurebase.
