LocalIntelligence

LocalIntelligence - Your Private AI Assistant with MCP support LocalIntelligence is a powerful application that brings the power of large language models directly to your desktop through Ollama. Chat with AI models privately and securely—everything runs locally on your device, with no cloud dependency or data sharing. Core Features Ollama Integration Connect seamlessly to your local Ollama instance. Browse and select from all your installed models, with automatic detection of model capabilities including vision, tools, thinking, and embedding support. Each model displays detailed runtime information including quantization level, context window size, and default generation parameters. Vision Support Vision-capable models can analyze images. Simply drag and drop an image onto the photo icon or use the file picker to attach photos, screenshots, or graphics. Ask questions about what you see, extract text, analyze diagrams, or get creative descriptions—all processed locally on your device. MCP (Model Context Protocol) Integration Extend your AI's capabilities with MCP tools. Connect to TCP/IP MCP servers. Tools are automatically discovered and made available to compatible models, enabling your AI to fetch real-time data, perform calculations, access APIs, and interact with external services. The app handles tool execution, result processing, and multi-turn conversations seamlessly. Fine-Tuned Control Adjust generation parameters on the fly with an intuitive interface: • Temperature: Control randomness and creativity (0.0 - 2.0) • Top P: Nucleus sampling for vocabulary focus • Top K: Limit token choices for consistency • Seed: Set fixed seeds for reproducible outputs Performance Monitoring Track model performance with detailed metrics (optional verbose mode): • Prompt evaluation speed (tokens/second) • Generation speed (tokens/second) • Context window usage with visual indicator • Load times and total duration Conversation Management • Export conversations in plain text or Markdown format • Copy to clipboard or save as files • Clear conversations with confirmation dialogs • Context window tracking prevents truncation issues • Model switching automatically clears conversations to prevent confusion Modern, native experience: • Clean, intuitive interface • Dark mode support • Keyboard shortcuts and native controls • Smooth animations and transitions • Efficient memory usage Privacy & Security • 100% local processing - no data leaves your Mac • No telemetry or tracking • Your conversations stay private Advanced Features Model Information View comprehensive details about each model: • Quantization type with explanatory guide (Q4_K_M, Q8_0, F16, etc.) • Current and maximum context window sizes • Parameter family and format • Size and modification date Context Window Management Visual indicator shows real-time context usage: • Automatic warnings when limits are reached MCP Server Management Configure and test MCP connections: • Connect to remote HTTP/SSE servers • Real-time connection testing with detailed diagnostics • Visual status indicators (connected, failed, testing) • Automatic tool discovery and listing Tool Execution When models with tool support call functions: • Visual indication of executing tools • Display of tool results inline • Multi-turn conversations with context preservation System Requirements • Ollama installed and running locally • Recommended: Apple Silicon for optimal performance Perfect For • Developers testing and debugging models • Researchers exploring AI capabilities • Privacy-conscious users who want local AI • Anyone who wants full control over their AI interactions LocalIntelligence puts powerful AI tools in your hands while respecting your privacy and giving you complete control over the experience.