Skip to main content

How to run a local Copilot

With Ollama and Continue​

Install ollama​

brew install ollama

This will install ollama using Homebrew, making it easy to upgrade by running brew upgrade ollama.

Start ollama server​

ollama serve

This will start ollama server.

You can also run ollama as a background service that will restart at login by running the following command instead.

brew services start ollama

Download codellama model weights​

ollama pull codellama:latest

You can also try other models listed here

Download and install Continue VSCode extension​

→ Visit https://continue.dev and follow instructions

Configure Continue​

Edit ~/.continue/config.py to replace default model by code-llama

# ~/.continue/config.py
from continuedev.src.continuedev.libs.llm.ollama import Ollama

config = ContinueConfig(
    ...
    models=Models(
        default=Ollama(model="codellama")
    )
)

See also: https://continue.dev/docs/walkthroughs/codellama

Set up VSCode​

Open VSCode, hit ´Cmd+Shift+P´ and select ´Continue: Focus on view´.

  • You can now select code + comments, and type ´/edit´ in Continue sidebar to start a completion.

Other setups​

  • Cody : Work in Progress