VS Code Extension

Luntrex Copilot

Automation-first, AI code suggestions directly inside VS Code. No chat UI, just fast inline completions powered by your Luntrex API key.

Get the extension Setup steps

Why you’ll like it

  • Inline, context-aware completions (no chat noise)
  • Works with any Luntrex completion model you’ve enabled
  • Respects your rate limits and API key security
  • Debounced requests for snappy typing
  • Excludes sensitive files by default (.env, keys, node_modules/vendor)

Setup (90 seconds)

  1. Create or sign in to your Luntrex account.
  2. Generate an API key in your dashboard.
  3. Install the VSIX (or marketplace build) of Luntrex Copilot.
  4. Open VS Code Settings → search “Luntrex Copilot” → paste your API key.
  5. Reload VS Code; verify status bar shows “Luntrex Copilot: On”.

Recommended settings

API Key

Paste your Luntrex API key.

Max Tokens

96 (safe default for short inline suggestions)

Debounce

120 ms (reduces extra calls while typing)

Exclude Globs

**/.env, **/.env.*, **/node_modules/**, **/vendor/**, **/dist/**, **/build/**, **/*.key

{
  "luntrexCopilot.apiKey": "sk_***",
  "luntrexCopilot.endpoint": "https://luntrex.com/api/v1/vscode/completions",
  "luntrexCopilot.model": "openai/gpt-4o-mini",
  "luntrexCopilot.maxTokens": 96,
  "luntrexCopilot.debounceMs": 120,
  "luntrexCopilot.excludeGlobs": [
    "**/.env", "**/.env.*",
    "**/node_modules/**", "**/vendor/**",
    "**/dist/**", "**/build/**",
    "**/*.key"
  ],
  "luntrexCopilot.enabled": true
}

How it works

The extension sends a trimmed window of code around your cursor, plus language info, to the dedicated /api/v1/vscode/completions endpoint. It returns code-only suggestions (no chat prose) as ghost text. Requests are debounced and aborted on new keystrokes to keep typing latency low.

Privacy: excluded globs prevent sending secrets; data travels over TLS; nothing is stored in the extension.

Back to top