Local AI

AI chat for your code, powered by a local Ollama LLM — fully offline, no cloud APIs, no data leaving your machine.

Features

💬 Ask

• Chat with AI about your code
• Streaming responses — tokens appear as the model generates them
• Markdown rendering — formatted output with syntax-highlighted code blocks
• Copy / Apply / Save buttons on every code block

📎 Context Attach

• Auto-context — active file is automatically attached when no text is selected
• Selection context — highlight code in the editor → auto-appears as context chip
• 📎 button — attach current file, selection, a folder (recursive), or browse individual files
• Drag & drop — drag files from Explorer or Finder directly into the chat
• Right-click — right-click any file/folder in Explorer → Add to AI Chat Context

🗂️ Conversation History

• Conversations are saved per workspace
• Use the 🕐 button to browse, switch, or delete past chats
• Auto-titled from the first message

🔧 Model Management

• Dropdown shows installed models with parameter size and quantization badges (e.g. 3.2B · Q4_K_M)
• Library models show size info and description before you pull
• One-click ↓ pull for any model with live progress bar
• Hover over installed model → ✕ to uninstall (two-step confirmation)
• Connection indicator: green = Ollama running, red = offline

Keyboard Shortcuts

Prerequisites

1. Install Ollama

# macOS / Linux
curl -fsSL https://ollama.com/install.sh | sh

# Windows / macOS — download from https://ollama.com/download

⚠️ Do not use brew install ollama — the Homebrew package is missing the llama-server binary.

2. Pull a model

# Lightweight & fast (good for low-resource machines)
ollama pull gemma:2b
ollama pull llama3.2:3b
ollama pull phi3:mini

# Coding-focused
ollama pull qwen2.5-coder:7b
ollama pull codellama:7b
ollama pull deepseek-coder:6.7b

# Balanced quality / speed
ollama pull mistral:7b
ollama pull llama3.1:8b

3. Start Ollama

ollama serve
# "address already in use" → Ollama is already running, that's fine

Configuration

Open Settings (Cmd+,) and search for localAIPrompt:

Run Locally (Development)

cd local-ai-prompt
npm install
npm run compile
# Press F5 in VS Code → opens Extension Development Host

Package as VSIX

npm install -g @vscode/vsce
vsce package

Install: Extensions → ... → Install from VSIX…

Troubleshooting

License

MIT