Cursor-Inline

Cursor-style inline AI editing for Neovim. Select code, describe the change, and get an inline, highlighted edit you can accept or reject—similar to Cursor’s inline workflow.

https://github.com/user-attachments/assets/55e2a362-19bf-4813-a734-ca28a9916b16

Features

• Inline popup for AI edits, triggered from visual selection.
• One-key accept or reject for generated inline edits.

Requirements

• Neovim with support for vim.system (0.10+ is recommended).
• curl available in your PATH.
• An OpenAI API key with access to the configured model (default: gpt-5.4-mini).

Providers Available

1. OpenAI
2. Anthropic

Installation

Use your favorite plugin manager. Examples below assume the repository path is nishu-murmu/cursor-inline – adjust if your repo is named differently.

lazy.nvim

{
  "nishu-murmu/cursor-inline",
  event = "BufReadPost",
  config = function()
    require("cursor-inline").setup()
  end,
}

packer.nvim

use({
  "nishu-murmu/cursor-inline",
  config = function()
    require("cursor-inline").setup()
  end,
})

vim-plug

Plug 'nishu-murmu/cursor-inline'

Then, somewhere in your Neovim config:

require("cursor-inline").setup()

Configuration

You can customize key mappings and the OpenAI model/provider via setup.

Default configuration from lua/cursor-inline/config.lua:

{
  mappings = {
    open_input = "<Space>e",
    accept_response = "<Space>y",
    deny_response = "<Space>n",
  },
  provider = {
    name = "openai",
    model = "gpt-5.4-mini",
  },
  preview = {
    diff = true,
  },
  streaming = {
    enabled = true,
  },
}

preview.diff enables the inline diff-style preview: generated code is highlighted as added code and the original selected code is highlighted as removed code until you accept or reject the edit.

streaming.enabled streams provider output into the inline preview as it is generated. Set it to false to wait for the full provider response before showing the edit.

---

TODOs

• Integrate multiple AI providers
• Adding spinner animations
• Diff previews for edits
• Streaming response support

---

License

MIT