LocalSeek 🔍✨

Seek your answers 💯% LOCALLY within VSCode

LocalSeek is a powerful, privacy-first AI chat extension for Visual Studio Code that brings conversational AI directly to your development environment - completely locally. Chat with your code, leverage your knowledge base, and get AI assistance without ever leaving your editor or compromising your privacy.

🌐 WEBSITE

🚀 What You Get

🤖 Local AI Chat

• Chat with AI models through Ollama without sending your data anywhere
• Choose between sidebar panel or standalone window
• Watch responses stream in real-time with full markdown rendering and syntax highlighting
• Switch between different models instantly

📖 RAG Integration

• Index your files to give AI context about your project
• Toggle "Use RAG" on/off per query
• AI automatically searches your Knowledge Base to provide relevant, project-specific answers instead of generic responses

💻 Seamless Code Workflow

• Select any code → right-click → "Send to LocalSeek Chat" for instant context
• AI responds with code? Click "Insert" to put it directly in your editor at cursor position
• Copy code blocks with one click
• All code gets proper syntax highlighting

💬 Smart Conversation Management

• All chats automatically saved with generated titles
• Resume any previous conversation exactly where you left off
• Browse your chat history with timestamps and message counts
• Clean up conversations you don't need anymore

🎛️ Built-in Model Manager - OLLAMA

• Download new Ollama models directly from the extension interface
• Watch real-time download progress with detailed status updates
• View model information like size and modification dates
• Remove unused models to free up disk space

🎨 VSCode-Native Interface

• Dark theme that matches your editor perfectly
• Responsive design that works on any screen size
• Smooth animations and intuitive controls
• Everything feels native to VSCode - no jarring external interfaces

🚀 Getting Started

Prerequisites

System Requirements

• Visual Studio Code (latest version recommended)
• Ollama installed and running locally
• Minimum 8GB RAM (16GB+ recommended for larger models)
• Available Storage for AI models (varies by model size)

Install Ollama

# Install Ollama (visit https://ollama.com for platform-specific instructions)
# Then pull some recommended models:
ollama pull gpt-oss            # Best model now
ollama pull deepseek-r1:14b    # Excellent reasoning model
ollama pull llama3.2:latest    # Versatile and reliable
ollama pull phi3:mini          # Lightweight and fast
ollama pull mistral:latest     # Great for coding tasks
ollama pull qwen2.5-coder      # Specialized for code generation

Installation

Method 1: VSCode Marketplace (Recommended)

1. Open VSCode
2. Go to Extensions (Ctrl+Shift+X / Cmd+Shift+X)
3. Search for "LocalSeek"
4. Click "Install"

Method 2: Manual Installation

1. Download the latest .vsix file from GitHub Releases
2. Open VSCode Extensions view
3. Click the "..." menu → "Install from VSIX"
4. Select the downloaded file

📖 Usage Guide

Opening LocalSeek

Sidebar Panel (Recommended)

• Click the LocalSeek icon in the Activity Bar (left sidebar)
• The chat panel opens in the sidebar for easy access while coding

Standalone Window

• Open Command Palette (Ctrl+Shift+P / Cmd+Shift+P)
• Type "LocalSeek: Open AI Chat"
• Chat opens in a separate panel

Basic Chat

1. Select Model: Choose your preferred Ollama model from the dropdown
2. Enable Knowledge Base: Toggle the "Use RAG" switch to include your indexed documents (off by default)
3. Type Message: Enter your question or request
4. Send: Press Enter or click the Send button
5. View Response: Watch the AI response stream in real-time

Working with Code

Send Code to Chat

1. Select any code in your editor
2. Right-click → "Send Selected Code to LocalSeek Chat"
3. The code appears in your chat input with proper formatting
4. Add your question and send

Insert AI Code

1. Click "Insert" button on any code block in AI responses
2. Code is automatically inserted at your cursor position
3. Replaces selected text if you have a selection

Knowledge Base Setup

1. Configure Knowledge Base Path (Required)

   • Open Settings (Ctrl+, / Cmd+,) and search "LocalSeek"
   • Set "Knowledge Base Path" to a specific directory - this is required to use RAG

2. Index Your Documents

   • Open Command Palette (Ctrl+Shift+P)
   • Type "LocalSeek: Index Knowledge Base"
   • Extension scans for files in the specified path

3. Use in Chat

   • Toggle the "Use RAG" switch on in the chat interface (off by default)
   • Ask questions related to your documentation
   • AI will automatically include relevant context from indexed files

Managing Conversations

View Chat History

• Click the history button (clock icon) in the chat interface
• Browse all your previous conversations
• Click any conversation to resume it

Start New Chat

• Click the new chat button (+ icon)
• Starts a fresh conversation
• Previous chat is automatically saved

Model Management

Download New Models

1. Click the models button (layers icon) in chat interface
2. Enter model name (e.g., "llama3.2", "deepseek-r1:7b")
3. Click "Download"
4. Monitor download progress in real-time

Remove Models

1. Open Model Management modal
2. Click "Remove" next to any installed model
3. Confirm deletion to free up disk space

🔧 Advanced Usage

Command Palette Commands

• LocalSeek: Open AI Chat - Open standalone chat window
• LocalSeek: Send Selected Code - Send selected code to chat
• LocalSeek: Index Knowledge Base - Index workspace documents

Tips & Tricks

1. Context Management: Use the KB toggle strategically - turn it off for general questions, on for project-specific queries

2. Model Selection:

   • Use smaller models (phi3) for quick questions
   • Use larger models (deepseek-r1) for complex reasoning
   • Use code-specific models for programming tasks

3. Efficient Workflows:

   • Keep sidebar chat open while coding
   • Use "Send Selected Code" for quick code reviews
   • Leverage chat history to build on previous conversations

4. Performance Optimization:

   • Close unused models to free RAM
   • Index only essential documents for faster search
   • Use smaller models for better response times

🔒 Privacy & Security

Privacy Guarantees

✅ 100% Local Processing - All AI inference happens on your machine ✅ No Data Transmission - Your code and conversations never leave your computer ✅ No Telemetry - Zero tracking or analytics ✅ Offline Capable - Works completely without internet connection ✅ Your Data, Your Control - Full ownership of all conversations and data

Security Features

• No External Dependencies for AI processing
• Local Storage Only for chat history and settings
• No API Keys Required - no risk of key exposure
• Open Source - transparent and auditable code

🤝 Contributing

We welcome contributions from the community!

Ways to Contribute

• 🐛 Report Bugs - Help us identify and fix issues
• 💡 Suggest Features - Share ideas for new functionality
• 📖 Improve Docs - Help make documentation clearer
• 🛠️ Submit Code - Contribute bug fixes or new features
• ⭐ Star the Repo - Show your support

🌟 What's Next?

Roadmap

• 📁 More File Types - Support for additional document formats
• 🔍 Advanced Search - Enhanced knowledge base search capabilities
• 🎨 Theme Customization - Multiple UI themes and customization options
• 🔌 Plugin System - Extensible architecture for custom integrations
• 📊 Analytics Dashboard - Usage insights and conversation analytics
• 🌐 Multi-language Support - Interface localization

Version History

Check the Changelog for detailed version history and updates.

📄 License

MIT License - see LICENSE file for details.

Developed with ❤️ by Hariharen

LocalSeek - Your local AI companion for VSCode. Seek your answers, locally and privately. 🔍✨