CodexComplete

CodexComplete is a VS Code autocomplete extension that delivers Copilot-style inline suggestions using OpenAI API models (default: gpt-5.3-codex).

Features (POC)

• Inline ghost-text autocomplete while typing.
• Manual command to complete at the current cursor.
• Fast defaults for low-latency coding flow.
• API key stored securely in VS Code SecretStorage.
• Activity Bar icon with a full CodexComplete GUI sidebar.
• In-panel settings editor (model, timeout, debounce, context, inline toggle).
• Model dropdown auto-loads all models available to the user's OpenAI API key.
• Live diagnostics + token usage charts grouped by day/week/month.
• Optional daily token limit (null = no limit).

Setup

1. Install dependencies:

   npm install
   

2. Build:

   npm run compile
   

3. Launch Extension Development Host from VS Code (Run Extension).
4. Run command: CodexComplete: Set OpenAI API Key.
5. Open the CodexComplete icon in the Activity Bar to use the GUI.

Commands

• CodexComplete: Set OpenAI API Key
• CodexComplete: Complete Now
• CodexComplete: Open Diagnostics Panel

Settings

• codexComplete.model
• codexComplete.requestTimeoutMs
• codexComplete.debounceMs
• codexComplete.maxContextChars
• codexComplete.enableInline
• codexComplete.dailyTokenLimit

Notes

• API key is never stored in plaintext settings.

Open Source

• License: MIT
• Contributing guide: CONTRIBUTING.md
• Community guidelines: CODE_OF_CONDUCT.md