CodeCompanion.nvim

CodeCompanion.nvim requires you to set your OpenAI API key as an environment variable and then configure the plugin in your init.lua or init.vim to reference it, along with your preferred model and keybindings. The plugin's setup function lets you customize completion triggers and the specific LLM backend to use.

Traditional completion engines like nvim-cmp use Language Server Protocol responses and snippet engines to suggest symbols and function names deterministically, while CodeCompanion.nvim leverages large language models to generate context-aware multi-line code suggestions. The LLM approach can handle open-ended prompts like 'write a function that parses CSV' rather than just completing the current identifier.

While the plugin was originally built around OpenAI's Codex and GPT models, many Neovim AI coding plugins have evolved to support alternative providers through configurable API endpoints. Check the plugin's current README and documentation for the most up-to-date list of supported backends.

Yes, the plugin provides inline code generation directly within the active buffer, so you can accept, reject, or edit AI-suggested code snippets without switching to a browser or external chat interface. This keeps the entire coding workflow inside the terminal for users who live in Neovim.