Nvidia has released an early version of Chat with RTX. Users can use this demo software to run a personalized AI chatbot on their PC. The chatbot was announced earlier this year during the CES trade fair in Las Vegas.
Chat with RTX can be downloaded immediately from Nvidia’s website, the company confirms in a blog post. The software is available for users with an RTX 30 or RTX 40 GPU with at least 8GB of video memory. The user’s system must also have at least 16GB of RAM and run Windows 11.
Nvidia already announced Chat with RTX in January. The company then said the software would be available later that month, but that ended up being mid-February. The tool is based on a large language model from Nvidia, which users can run locally. They can ‘connect’ the language model with their own data, after which the chatbot can answer questions about it.
For example, users can provide the software with locally stored text documents, PDFs and XML files. It is also possible to enter URLs of YouTube videos and playlists into the software, and the software will download transcripts of those videos. The bot can analyze that various data and then users can ask questions about it. The demo runs locally and uses the Tensor cores in the RTX GPU.