Local where possible
A project focused on fast answers, affordable running costs and greater control over your environment.
A project to make artificial intelligence feel closer, more useful and more personal.
Vega is an AI assistant for Windows PCs in development. It brings together local conversation, voice, commands and practical tools. Here we separate current features from prototypes and open problems.
A project focused on fast answers, affordable running costs and greater control over your environment.
Working features, prototypes and limitations will be explained clearly, without premature promises.
We welcome people interested in contributing code, testing and a better user experience.
The Windows program exists and is developed and tested locally. These capabilities are part of the project; quality varies by task and environment.
Speak or type through the same request pipeline. Conversation runs on Qwen3.5:4B through Ollama on the PC.
Local commands, a countdown timer and web search in a dedicated window.
Travel and visit details can be extracted, stored and corrected with dedicated checks. Open conversation remains imperfect.
Particle face, iris standby and dedicated windows, including mail. Real behavior depends on the PC and external services.
Vega is not a finished product. We are working to reduce repetition, wrong attributions and invented answers in open conversation. Some promising results are separate prototypes, not features in the program yet.
We welcome contributors in coding, Windows testing, issue analysis, accessibility, documentation and interface quality.
Direct chat between the site and Vega is being designed. When ready, a request will reach the creator only while connected and available, and they can accept it. Chat is not active yet. If we are offline, email us and we will reply.