AI and Copilot Plugins

AI-assisted coding has become a standard part of modern development workflows. Neovim has best-in-class integrations for all major AI coding tools.

1. GitHub Copilot — copilot.lua (Official-quality)

The fastest Copilot integration — async, native Lua, no Node.js overhead in the editor:

lua/plugins/ai.lua
return {
  -- Core Copilot engine
  {
    "zbirenbaum/copilot.lua",
    cmd = "Copilot",
    event = "InsertEnter",
    config = function()
      require("copilot").setup({
        panel = {
          enabled = true,
          auto_refresh = false,
          keymap = {
            jump_prev = "[[",
            jump_next = "]]",
            accept    = "<CR>",
            refresh   = "gr",
            open      = "<M-CR>",
          },
          layout = {
            position = "bottom",  -- top | left | right | bottom
            ratio    = 0.4,
          },
        },
        suggestion = {
          enabled = true,
          auto_trigger = true,        -- show suggestions automatically
          hide_during_completion = true,
          debounce = 75,
          keymap = {
            accept        = "<M-l>",   -- Alt+l to accept
            accept_word   = false,
            accept_line   = "<M-L>",   -- Alt+Shift+l to accept line
            next          = "<M-]>",   -- Alt+] next suggestion
            prev          = "<M-[>",   -- Alt+[ prev suggestion
            dismiss       = "<C-]>",   -- Ctrl+] dismiss
          },
        },
        filetypes = {
          yaml        = false,
          markdown    = false,
          help        = false,
          gitcommit   = false,
          gitrebase   = false,
          ["."]       = false,         -- disable for dotfiles by default
          -- Explicitly enable for:
          lua         = true,
          python      = true,
          javascript  = true,
          typescript  = true,
          php         = true,
          bash        = true,
          sh          = true,
        },
        copilot_node_command = "node",  -- or path to node binary
        server_opts_overrides = {},
      })
    end,
  },

  -- Copilot as nvim-cmp source (optional: shows as completion items)
  {
    "zbirenbaum/copilot-cmp",
    dependencies = { "zbirenbaum/copilot.lua" },
    config = function()
      require("copilot_cmp").setup()
    end,
  },
}

Authentication

# First time setup
nvim
:Copilot setup     → opens browser to authenticate
:Copilot status    → verify it's enabled
:Copilot enable    → enable for current session
:Copilot disable   → disable for current session

Key Bindings

Key	Action
`Alt+l`	Accept full suggestion
`Alt+Shift+l`	Accept one line
`Alt+]`	Next suggestion
`Alt+[`	Previous suggestion
`Ctrl+]`	Dismiss suggestion

2. Avante.nvim — Cursor-style AI Chat in Neovim

Avante is an AI chat panel that reads your code, understands context, and can directly apply AI-suggested changes to your buffer:

{
  "yetone/avante.nvim",
  event = "VeryLazy",
  version = false,
  build = "make",    -- builds native modules for performance
  dependencies = {
    "nvim-treesitter/nvim-treesitter",
    "stevearc/dressing.nvim",
    "nvim-lua/plenary.nvim",
    "MunifTanjim/nui.nvim",
    "nvim-tree/nvim-web-devicons",
    "zbirenbaum/copilot.lua",  -- optional
    {
      "HakonHarnes/img-clip.nvim",  -- paste images into chat
      event = "VeryLazy",
      opts = {
        default = {
          embed_image_as_base64 = false,
          prompt_for_file_name = false,
          drag_and_drop = { insert_mode = true },
          use_absolute_path = true,
        },
      },
    },
    {
      "MeanderingProgrammer/render-markdown.nvim",  -- render markdown in buffer
      opts = { file_types = { "markdown", "Avante" } },
      ft = { "markdown", "Avante" },
    },
  },
  opts = {
    -- Choose AI provider:
    provider = "copilot",       -- copilot | claude | openai | azure | gemini | ollama

    -- Provider configs
    copilot = {
      model = "gpt-4o",
    },
    openai = {
      endpoint = "https://api.openai.com/v1",
      model = "gpt-4o",
      timeout = 30000,
      temperature = 0,
      max_tokens = 4096,
    },
    claude = {
      endpoint = "https://api.anthropic.com",
      model = "claude-3-7-sonnet-20250219",
      timeout = 30000,
      temperature = 0,
      max_tokens = 8000,
    },
    ollama = {
      endpoint = "http://127.0.0.1:11434",
      model = "llama3.1",
    },

    -- Behaviour
    behaviour = {
      auto_suggestions = false,        -- experimental
      auto_set_highlight_group = true,
      auto_set_keymaps = true,
      auto_apply_diff_after_generation = false,
      support_paste_from_clipboard = false,
    },

    -- Keymaps
    mappings = {
      diff = {
        ours   = "co",
        theirs = "ct",
        all_theirs = "ca",
        both   = "cb",
        cursor = "cc",
        next   = "]x",
        prev   = "[x",
      },
      suggestion = {
        accept_all = "<M-a>",
        accept_cursor = "<M-l>",
        reject_all = "<M-r>",
      },
      jump = { next = "]]", prev = "[[" },
      submit = { normal = "<CR>", insert = "<C-s>" },
      ask = "<leader>aa",
      edit = "<leader>ae",
      refresh = "<leader>ar",
      focus = "<leader>af",
      toggle = {
        default = "<leader>at",
        debug   = "<leader>ad",
        hint    = "<leader>ah",
        suggestion = "<leader>as",
        repomap = "<leader>aR",
      },
      files = {
        add_current = "<leader>ac",
        add_all_buffers = "<leader>aA",
      },
    },
  },
},

Avante Workflow

<leader>aa   → open Avante chat panel with current file context
              → ask: "Refactor this function to be more readable"
              → Avante shows a diff of suggested changes
              → press co/ct to accept/reject hunks (like git merge)

<leader>ae   → ask about selected code (visual mode)
<leader>at   → toggle the Avante panel

3. Codeium — Free GitHub Copilot Alternative

Codeium is a free AI completion tool:

{
  "Exafunction/codeium.nvim",
  dependencies = {
    "nvim-lua/plenary.nvim",
    "hrsh7th/nvim-cmp",
  },
  event = "InsertEnter",
  config = function()
    require("codeium").setup({
      enable_chat = false,
    })
    -- Add to cmp sources:
    -- { name = "codeium", group_index = 1, priority = 100 },
  end,
},

Setup: :Codeium Auth → paste your API key from codeium.com (free account).

4. gen.nvim — Ollama Local AI

Use local LLM models with Ollama, no internet required:

{
  "David-Kunz/gen.nvim",
  cmd = "Gen",
  keys = {
    { "<leader>ai", "<cmd>Gen<cr>",               mode = { "n", "v" }, desc = "Open AI (gen.nvim)" },
    { "<leader>ac", "<cmd>Gen Chat<cr>",           desc = "AI Chat" },
    { "<leader>ag", "<cmd>Gen Generate<cr>",       desc = "AI Generate" },
    { "<leader>as", "<cmd>Gen Summarize<cr>",      mode = "v",          desc = "AI Summarize" },
  },
  opts = {
    model = "llama3.1",        -- your Ollama model
    host = "localhost",
    port = "11434",
    display_mode = "float",    -- float | split
    show_prompt = true,
    show_model = true,
    no_auto_close = false,
    quit_map = "q",
    retry_map = "<c-r>",
    accept_map = "<c-cr>",
    init = function(options)
      pcall(io.popen, "ollama serve > /dev/null 2>&1 &")
    end,
    command = function(options)
      return "curl --silent --no-buffer -X POST http://" ..
        options.host .. ":" .. options.port .. "/api/chat -d $body"
    end,
    result_filetype = "markdown",
    debug = false,
  },
},

Prerequisites

# Install Ollama
curl -fsSL https://ollama.com/install.sh | sh

# Pull a model
ollama pull llama3.1
ollama pull codellama
ollama pull deepseek-coder-v2

# Verify
ollama list

5. ChatGPT.nvim — Direct ChatGPT Integration

{
  "jackMort/ChatGPT.nvim",
  event = "VeryLazy",
  dependencies = {
    "MunifTanjim/nui.nvim",
    "nvim-lua/plenary.nvim",
    "folke/trouble.nvim",
    "nvim-telescope/telescope.nvim",
  },
  config = function()
    require("chatgpt").setup({
      api_key_cmd = "cat ~/.config/openai_key",   -- read key from file
      -- or: api_key_cmd = "pass openai/api-key"  -- from password manager
      yank_register = "+",
      extra_curl_params = {},
      openai_params = {
        model = "gpt-4o",
        frequency_penalty = 0,
        presence_penalty = 0,
        max_tokens = 300,
        temperature = 0,
        top_p = 1,
        n = 1,
      },
      openai_edit_params = {
        model = "gpt-4o",
        frequency_penalty = 0,
        presence_penalty = 0,
        temperature = 0,
        top_p = 1,
        n = 1,
      },
      use_openai_functions_for_edits = false,
      actions_paths = {},
      show_quickfixes_cmd = "Trouble quickfix",
      predefined_chat_gpt_prompts = "https://raw.githubusercontent.com/f/awesome-chatgpt-prompts/main/prompts.csv",
      keymaps = {
        close                  = "<C-c>",
        yank_last              = "<C-y>",
        yank_last_code         = "<C-k>",
        scroll_up              = "<C-u>",
        scroll_down            = "<C-d>",
        new_session            = "<C-n>",
        cycle_windows          = "<Tab>",
        cycle_modes            = "<C-f>",
        next_message           = "<C-j>",
        prev_message           = "<C-k>",
        select_session         = "<Space>",
        rename_session         = "r",
        delete_session         = "d",
        draft_message          = "<C-r>",
        edit_message           = "e",
        delete_message         = "d",
        toggle_settings        = "<C-o>",
        toggle_sessions        = "<C-p>",
        toggle_help            = "<C-h>",
        toggle_message_role    = "<C-r>",
        toggle_system_role_open = "<C-s>",
        stop_generating        = "<C-x>",
      },
    })

    vim.keymap.set("n", "<leader>ac", "<cmd>ChatGPT<CR>", { desc = "ChatGPT" })
    vim.keymap.set({"n","v"}, "<leader>ae", "<cmd>ChatGPTEditWithInstruction<CR>", { desc = "Edit with AI" })
    vim.keymap.set({"n","v"}, "<leader>ag", "<cmd>ChatGPTRun grammar_correction<CR>", { desc = "Grammar fix" })
    vim.keymap.set({"n","v"}, "<leader>at", "<cmd>ChatGPTRun translate<CR>", { desc = "AI Translate" })
    vim.keymap.set({"n","v"}, "<leader>aD", "<cmd>ChatGPTRun docstring<CR>", { desc = "Add docstring" })
    vim.keymap.set({"n","v"}, "<leader>aA", "<cmd>ChatGPTRun add_tests<CR>", { desc = "Add tests" })
    vim.keymap.set({"n","v"}, "<leader>ao", "<cmd>ChatGPTRun optimize_code<CR>", { desc = "Optimize code" })
    vim.keymap.set({"n","v"}, "<leader>aS", "<cmd>ChatGPTRun summarize<CR>", { desc = "Summarize" })
    vim.keymap.set({"n","v"}, "<leader>af", "<cmd>ChatGPTRun fix_bugs<CR>", { desc = "Fix bugs" })
    vim.keymap.set({"n","v"}, "<leader>ax", "<cmd>ChatGPTRun explain_code<CR>", { desc = "Explain code" })
  end,
},

AI Plugin Comparison

Plugin	Provider	Cost	Offline	Code Apply
`copilot.lua`	GitHub Copilot	$10/mo	❌	Via suggestion
`avante.nvim`	Copilot/Claude/OpenAI	Varies	Via Ollama	✅ Diff apply
`codeium.nvim`	Codeium	Free	❌	Via suggestion
`gen.nvim`	Ollama (local)	Free	✅	Manual
`ChatGPT.nvim`	OpenAI	Pay-per-use	❌	Via instruction

1. GitHub Copilot — copilot.lua (Official-quality)​

Authentication​

Key Bindings​

2. Avante.nvim — Cursor-style AI Chat in Neovim​

Avante Workflow​

3. Codeium — Free GitHub Copilot Alternative​

4. gen.nvim — Ollama Local AI​

Prerequisites​

5. ChatGPT.nvim — Direct ChatGPT Integration​

AI Plugin Comparison​

What's Next​