ollama-next

Talk to it with a chat model

Ask it code questions with a code model

Install

• Install the ollama cli
  • at 'https://ollama.com/download'
• Get the llama3.2:3b model
  • ollama pull llama3.2:3b
• Get the qwen2.5-coder:7b model
  • ollama pull qwen2.5-coder:7b
• Run the ollama model server
  • ollama serve
• Run the ollama-next model ui
  • pnpm install
  • pnpm run dev

Other useful commands

• ollama ps
• ollama run llama3.2:3b
• ollama stop llama3.2:3b

How it works

• You install and run the ollama server
• You pull the running model which the code desires
  • See src/server/trpc/router.ts where the model is ran and queried
  • There is a line which tells the server which model to use
• The ui is an interface to chat with the model
• It "remembers" the last 7 or so things you asked about to simulate memory