CodeContinue

Fast inline code autocomplete, real-time streaming chat, and in-place code editing for Sublime Text 4. Powered by any OpenAI-compatible or Anthropic LLM endpoint.

Features

• ⚡ Fast Inline Autocomplete: Fill-In-The-Middle (FIM) and bidirectional context slicing with automatic suffix deduplication.
• 💬 Real-Time Streaming Chat: Discuss code in an interactive split Markdown panel with live token streaming (Server-Sent Events).
• ✏️ In-Place Edit & Diff Review: Rewrite code using natural instructions. Inspect proposed changes as a unified diff before applying.
• 🔀 Dual-Engine Architecture: Mix and match models (e.g. fast local model for completions, Claude 3.5 Sonnet / GPT-4o for chat and edits).
• 🎯 1-Click Quick Action Presets: Right-click selected code to Explain Selection, Find Bugs, Generate Unit Tests, Add Docstrings & Types, or Optimize Code.
• 🌐 Universal Provider Support: Works out of the box with LM Studio, Ollama, vLLM, OpenAI, Anthropic, or any OpenAI-compatible server.

---

Quick Start

1. Install via Package Control

1. In Sublime Text, press Ctrl+Shift+P (or Cmd+Shift+P on macOS).
2. Select Package Control: Install Package.
3. Search for CodeContinue and press Enter.

2. Run the Setup Wizard

Open the Command Palette (Ctrl+Shift+P / Cmd+Shift+P) and run CodeContinue: Configure (or go to Preferences > Package Settings > CodeContinue > Configure Wizard...).

The setup wizard will:

1. Connect to your endpoint (e.g. http://localhost:1234 or https://api.openai.com).
2. Auto-detect available models from your server.
3. Install recommended keybindings with a single click.

3. Usage

• Suggest code: Press Ctrl+Enter at any cursor position.
• Accept suggestion: Press Tab (or Escape to dismiss).
• Chat or Edit: Select code and right-click to choose CodeContinue actions.
• Switch models: Press Ctrl+Shift+P >> CodeContinue: Switch Model... to change models on the fly.

---

Command Reference

Access these commands via Ctrl+Shift+P / Cmd+Shift+P or the editor right-click menu:

Command	Menu	Description
CodeContinue: Suggest	—	Trigger code suggestion at cursor (Ctrl+Enter)
CodeContinue: Dismiss Suggestion	—	Dismiss active suggestion (Escape)
CodeContinue: Configure	—	Launch interactive setup wizard
CodeContinue: Switch Model...	—	Switch active model on the fly
CodeContinue: Chat about Selection	✓	Open streaming chat about highlighted code
CodeContinue: Explain Selection	✓	Break down logic, edge cases, and complexity
CodeContinue: Find Bugs & Vulnerabilities	✓	Scan selection for bugs and security issues
CodeContinue: Generate Unit Tests	✓	Generate tests with edge-case coverage
CodeContinue: Edit Selection	✓	Rewrite selection with instruction & diff preview
CodeContinue: Add Docstrings & Types	✓	Add clean docstrings and type annotations
CodeContinue: Optimize & Clean Code	✓	Refactor code for readability and performance
Preferences: CodeContinue Settings	—	Open user settings editor
Preferences: CodeContinue Key Bindings	—	Open keybinding editor

---

Keybindings

The setup wizard (CodeContinue: Configure) automatically installs the recommended shortcuts into your User keymap without conflicts:

Action	Shortcut	Command	Condition
Suggest	Ctrl+Enter	code_continue_suggest	Always available
Accept	Tab	code_continue_accept	Only when suggestion is visible
Dismiss	Escape	code_continue_dismiss	Only when suggestion is visible

To customize your keybindings manually, run Preferences: CodeContinue Key Bindings.

---

Documentation & Guides

• ⚙️ Configuration Guide: Full settings reference, dual-engine setup, and provider recipes (LM Studio, Ollama, OpenAI, Anthropic).
• 🛠️ Troubleshooting Guide: Common fixes, debug logs ("debug": true), and resolving manual install overrides.
• 📦 Developer Tools & Offline Install: Standalone CLI and GUI installers for development builds.

---

Requirements

• Sublime Text 4 (Build 4000+)
• Python 3.8+ (bundled with Sublime Text 4)
• Access to an LLM runner (LM Studio, Ollama, vLLM) or API endpoint (OpenAI, Anthropic, etc.)

---

License

See LICENSE for details.