Example of using Vercel's useChat with Llama.cpp /completion route. Handles formatting the data to match the OpenAI spec along with the ability to configure Llama.cpp parameters.
npm install
LLAMA_SERVER_URL=https://huggingface.co/spaces/matthoffner/ggml-coding-cpu/completion npm run dev
LLAMA_SERVER_URL defaults to http://127.0.0.1:8080/completion