web-llm-embed

💎 Edge based document chat. No data is sent to the server. Model is Vicuna 7b trained by LMSys.
📄 This will download the model (~4GB) and may take a few minutes. After the first time, it will be cached.