ConvoCode Chat

A Visual Studio Code extension that lets you chat with Ollama models directly from your editor.

Features

• Send messages to any local or remote Ollama model
• Live streaming of responses with partial updates
• Markdown rendering in chat history
• Select from one or more configured models
• Easy build, debug, and package workflow

Prerequisites

• Node.js v14 or higher
• Visual Studio Code v1.70.0 or higher
• A running Ollama API server (by default: http://127.0.0.1:11434)

Installation

Clone the repo and install dependencies:

git clone https://github.com/yourname/convocode.git
cd convocode
npm install

Development

1. Compile TypeScript and watch for changes:

   npm run watch
   

2. Launch the Extension Development Host:

   • Open this folder in VS Code
   • Press F5 to launch a new window with ConvoCode loaded

3. In the development host, open the command palette (Ctrl+Shift+P) and run:

   ConvoCode Chat: Start Conversation
   

4. Type your messages into the input box and hit Send.

Configuration

Settings can be modified in your VS Code settings.json under the convocode namespace:

{
  "convocode.baseUrl": "http://127.0.0.1:11434",       // Ollama API endpoint
  "convocode.model": "qwen3:0.6b",                     // Default model
  "convocode.models": ["qwen3:0.6b", "other-model:tag"]  // Available models
}

Packaging & Publishing

1. Compile the extension:

   npm run compile
   

2. Create a VSIX package:

   npx vsce package
   

3. Install the generated .vsix locally:

   code --install-extension convocode-0.0.1.vsix
   

4. (Optional) Publish to the VS Code Marketplace:

   npx vsce publish
   

Release Notes

0.0.1

• Initial release with streaming chat, model selection, and markdown support.

Contributions, issues, and feature requests are welcome! Developed by Your Arun Kumar Tiwary