The main idea is to let you use the latest and most powerful AI models like GPT-4.1, Gemini 2.5 Pro, and Claude 4 without sacrificing your privacy. You can plug in your own API keys (BYOK) and chat directly with these models. There are also 9 free models available without needing a key.
Beyond simple chat, VT includes an intelligent tool router that can automatically use tools like document parsing and web search to answer complex queries. This allows the models to work with your private documents or fetch up-to-date information from the web, all while maintaining the local-first architecture.
The project is built with Next.js 14 and TypeScript, and the source is available on GitHub (https://github.com/vinhnx/vtchat) under the MIT license.
I'm looking for feedback from the HN community on the architecture and features. Thanks for checking it out.