GhostWriter

GhostWriter is a VS Code extension that provides AI-powered code assistance using Ollama models.

Features

• AI code generation and suggestions using local Ollama models
• Customizable system prompts to control the AI's behavior
• Chat interface similar to Cursor's AI assistant
• Code insertion directly into your editor
• Support for multiple Ollama models

Requirements

• Ollama must be installed and running locally
• A compatible AI model must be pulled in Ollama (e.g., llama3, codellama, etc.)

Development

Building the Extension

The GhostWriter extension includes a fully automated build script that intelligently determines the appropriate version increment based on your git commit messages:

./build.sh

Smart Version Management:

• Automatically analyzes git commit messages to determine version changes
• Major version (1.0.0 → 2.0.0): Triggered by commits with "BREAKING CHANGE" or "!:"
• Minor version (0.1.0 → 0.2.0): Triggered by commits with "feat:" or "feature:"
• Patch version (0.0.1 → 0.0.2): Default for bug fixes, documentation, etc.
• Release candidate (0.1.0 → 0.1.0-rc.1): Triggered by commits with "rc" or "release candidate"

Automatic Workflow:

1. Analyzes your commit history since the last tag
2. Determines the appropriate version increment
3. Updates package.json with the new version
4. Compiles and packages the extension
5. Creates a git tag for the new version
6. Keeps the current and previous build for reference

The script also manages release candidates, incrementing RC numbers as needed and finalizing them to stable versions when appropriate.

Installation

From VS Code Marketplace (Coming Soon)

Search for "GhostWriter" in the VS Code marketplace and click install.

From Source

1. Clone this repository
2. Run npm install to install dependencies
3. Run npm run compile to build the extension
4. Press F5 to launch a new VS Code window with the extension loaded

Publishing

This extension uses GitHub Actions to automatically publish new versions to the VS Code Marketplace when a new release is created.

Prerequisites for Publishing

1. Create a Personal Access Token (PAT) for the VS Code Marketplace:

   • Go to https://dev.azure.com/
   • Create or use an existing organization
   • Generate a Personal Access Token with the "Marketplace publishing" scope

2. Create a PAT for the Open VSX Registry (optional):

   • Go to https://open-vsx.org/
   • Create a namespace and token

3. Add the tokens as secrets in your GitHub repository:

   • Go to your repository's Settings > Secrets > Actions
   • Add VSCE_PAT with your VS Code Marketplace token
   • Add OPEN_VSX_PAT with your Open VSX token (optional)

Publishing Process

1. Create a new release on GitHub
2. The workflow will automatically:
   • Build and package the extension
   • Publish it to the VS Code Marketplace
   • Publish it to the Open VSX Registry (if configured)

You can also manually trigger the publishing workflow from the Actions tab and optionally specify a version.

Usage

1. Make sure Ollama is running locally with your preferred model pulled
2. Open the GhostWriter sidebar in VS Code
3. Start typing in the chat input to generate code or get programming assistance
4. Use the "Insert Code" button to add code snippets directly to your editor
5. Customize the system prompt and model using the GhostWriter commands

Available Commands

• GhostWriter: Start Chat - Opens the chat interface
• GhostWriter: Set AI Model - Change the Ollama model to use
• GhostWriter: Set System Prompt - Customize the system prompt

Extension Settings

This extension contributes the following settings:

• ghostwriter.ollamaUrl: URL for the Ollama API (default: http://localhost:11434)
• ghostwriter.model: Default model to use for code generation
• ghostwriter.systemPrompt: System prompt for the AI model

Troubleshooting

• If you see "Error: Could not generate response", make sure Ollama is running and the selected model is pulled
• Run ollama list in the terminal to see available models
• Run ollama pull llama3 (or other model name) to download a model