CodeContinue v2.5.0
===================

Major Features & Enhancements:

1. Real-Time Streaming Chat:
   • Responses stream token-by-token directly into the split Markdown chat view using
     Server-Sent Events (SSE). Works seamlessly with OpenAI and Anthropic models.

2. Dual-Engine Architecture:
   • Mix-and-match models: run a lightning-fast local model (e.g. Qwen 2.5 Coder 1.5B)
     for inline completions while routing Chat and In-Place Editing to Claude 3.5 Sonnet
     or GPT-4o. Configure chat_endpoint, chat_model, edit_endpoint, edit_model in settings.

3. Fill-In-The-Middle (FIM) & Suffix-Aware Completions:
   • The autocomplete engine now provides bidirectional prefix and suffix context with
     automatic suffix deduplication, ensuring clean insertions within existing code.

4. Interactive Diff & Review for "Edit Selection":
   • Refactored code can now be previewed as a unified diff in a dedicated scratch tab
     before accepting, preventing unwanted file modifications.

5. 1-Click Context Menu & Command Presets:
   • Right-click or press Ctrl+Shift+P to instantly:
     - Explain Selection (breaks down logic, edge cases, complexity)
     - Find Bugs & Vulnerabilities (identifies performance issues and bugs)
     - Generate Unit Tests (produces comprehensive tests with edge cases)
     - Add Docstrings & Types (annotates functions and classes inline)
     - Optimize & Clean Code (refactors for readability and performance)

6. Status Bar Widget & On-the-Fly Model Switcher:
   • Live status bar indicator shows the currently active model and completion activity.
   • Press Ctrl+Shift+P >> "CodeContinue: Switch Model..." to switch models on the fly.

7. Clean Defaults & Split-View Keymap Workflow:
   • Default endpoint set to standard http://localhost:8000/v1/chat/completions.
   • Step-by-step keybinding split-view guide prevents key conflict issues.

Questions or issues?
https://github.com/kXborg/codeContinue/issues
