Rift AI Coding Assistant
A VS Code extension that brings local AI coding assistance to your editor using Ollama. Your code stays on your machine — no cloud API required.
Install
- Install Ollama and start it
- Pull a model:
ollama pull llama3
- In VS Code, open Extensions (
Ctrl+Shift+X) and search for Rift AI Coding Assistant
- Click Install
Or install from the command line:
code --install-extension TewodrosSeble.rift-ai-coding-assistant
Quick start
- Open any code file in VS Code
- Press
Ctrl+Shift+P and type Rift AI
- Try Rift AI: Ask Local AI or select code → right-click → Rift AI: Explain Selected Code
- Click the Rift AI icon in the left activity bar (below Extensions) to open chat
Keyboard shortcut: Ctrl+Alt+R opens the Rift AI chat panel.
Don't see the icon? Reload VS Code (Ctrl+Shift+P → Developer: Reload Window), then scroll the left activity bar — extension icons appear at the bottom. Right-click the activity bar and ensure Rift AI is checked. Or press Ctrl+Alt+R to open chat directly.
Features
| Command |
What it does |
| Ask Local AI |
Ask any coding question |
| Explain Selected Code |
Plain-language explanation in a side panel |
| Refactor Selected Code |
Improve readability and structure in place |
| Generate Tests |
Create unit tests for selected code |
| Fix Error |
Fix code using a pasted error message |
| Add Comments |
Add helpful comments to selected code |
| Convert Code |
Convert selection to another language |
| Optimize Code |
Performance improvements in place |
| Index Workspace (RAG) |
Index your project for smarter context |
Chat sidebar — ongoing conversation in the activity bar.
Workspace context — optionally includes open files, package.json, and README in prompts.
Settings
Open VS Code Settings and search for localAiCoder:
| Setting |
Default |
Description |
localAiCoder.model |
llama3:latest |
Ollama model for chat and code tasks |
localAiCoder.ollamaUrl |
http://localhost:11434 |
Ollama server URL |
localAiCoder.embeddingModel |
nomic-embed-text |
Model for workspace indexing (RAG) |
localAiCoder.includeWorkspaceContext |
true |
Include project context in prompts |
localAiCoder.inlineCompletionEnabled |
false |
AI inline autocomplete (can be slow) |
Recommended models
ollama pull llama3 # general coding
ollama pull qwen2.5-coder:7b # code-focused
ollama pull nomic-embed-text # for RAG indexing (optional)
RAG (project-aware context)
For better answers about your codebase:
- Run
ollama pull nomic-embed-text
Ctrl+Shift+P → Local AI: Index Workspace (RAG)- Use Ask Local AI or the chat sidebar — relevant files are included automatically
Requirements
- VS Code 1.85 or newer
- Ollama running locally
- At least one Ollama model pulled
License
MIT
