Hover Explain

Hold Cmd (macOS) or Ctrl (Windows/Linux) and hover a variable, function or any other symbol. A short explanation of what it holds or does appears in the normal editor hover.

Use whichever model you like: NVIDIA's free API, OpenAI, OpenRouter, Groq, Google Gemini, a local model in Ollama or LM Studio, any other OpenAI-compatible API, or Claude Code.

Works in Cursor, VS Code, Windsurf, VSCodium and other VS Code–based editors, for every language that has a language server.

How to use

1. Hover a symbol and press Cmd / Ctrl, or hold the key first and then move onto the symbol.
2. Keep the mouse there; you can release the key. The hover shows "Loading…" and fills in the explanation when it arrives, labelled with the provider and model that wrote it.
3. Hovering the same symbol again shows the explanation straight away, until you edit the file.

You can also run Hover Explain: Explain Symbol at Cursor from the Command Palette to explain the symbol at the text cursor.

Setup

Run Hover Explain: Choose Provider from the Command Palette. It walks you through three steps:

1. Provider: pick where explanations come from.
2. API key: if the provider needs one, paste it. Keys are kept in the editor's secret storage, not in your settings file, and each provider has its own.
3. Model: pick from the provider's live model list, or type any model id.

To switch models later, run Hover Explain: Choose Model. To replace a key, run Hover Explain: Set API Key.

What it sends

When you trigger an explanation, the extension sends the chosen provider:

• the symbol's name, the file's relative path and its language
• about 25 lines on each side of the symbol
• about 10 lines on each side of where it is defined, if that is elsewhere
• up to 6 other lines where it is used

Nothing is sent when you hover without the modifier key. If your code must not leave your machine, use Ollama or LM Studio.

Settings

Commands

• Hover Explain: Choose Provider
• Hover Explain: Choose Model
• Hover Explain: Set API Key (submit an empty value to remove the key)
• Hover Explain: Explain Symbol at Cursor
• Hover Explain: Clear Cache

How the Cmd+hover trigger works

Extensions can't see which keys are held during a hover. While Cmd/Ctrl is held, though, the editor asks extensions for the definition of the word under the mouse so it can underline it. Hover Explain answers that request with nothing, so Go to Definition and Cmd+click work as before, and uses it as the trigger.

Definitions are also looked up in the background (by the editor's own AI features, for example). Those never cause a request: an explanation starts only when the editor also asks for a hover on that same word, which happens only for the word under the mouse.

If you were already resting on the word when you pressed Cmd, the editor has finished that hover and won't ask again until the mouse moves. Hover Explain then reopens the hover itself; the editor only allows that at the text cursor, so the cursor is moved onto the word for an instant and put straight back.

License

MIT