Skip to content

Releases: kXborg/codeContinue

CodeContinue with Inline edit

Pre-release

Choose a tag to compare

@kXborg kXborg released this 28 Jun 11:17

🚀 New Features

  • Inline Edit / Refactor (Ctrl+K style): You can now quickly refactor code right in your editor! Select a block of code, right-click, and choose CodeContinue: Edit Selection to instantly rewrite your code based on a custom instruction.
  • Anthropic / Messages API Support: You can now use Anthropic's Claude models seamlessly. Just configure an Anthropic endpoint (e.g., https://api.anthropic.com/v1/messages), an Anthropic model, and your API key in the plugin settings.

✨ Enhancements

  • Upgraded Chat UI: The selection chat view has been upgraded with richer Markdown formatting. It now uses proper Markdown headers, language-specific code blocks for the selected code, and visual separators (---) for a much cleaner and readable conversation flow.

Questions or issues? Feel free to open an issue here.

codeContinue SublimeText

Pre-release

Choose a tag to compare

@kXborg kXborg released this 25 May 03:16

An LLM-powered Sublime Text 4 plugin for inline code completion and chat-about-selection. Works with any OpenAI-compatible /v1/chat/completions endpoint like OpenAI, LMStudio, Ollama, vLLM, and others.

Fixes

  • API key now sent with completion requests — cloud endpoints (OpenAI, etc.) work
  • Default timeout fallback corrected (was 0.15s, now 30s)
  • HTML-escape phantom preview so <, >, & in completions render correctly (C++ templates, JSX)
  • First-run wizard prompts for an optional API key
  • Multi-line accept preserves the model's intended indentation
  • Wider model control-token cleanup (Llama 3, ChatML, FIM, GPT-style)
  • CodeContinue: Configure no longer crashes when run from the Command Palette
  • Errors visible in the Sublime console without enabling debug
  • Network/parse error messages reliably reach the status bar