Sep 14th, 2025
This week we focused on expanding 120 AI Chat local AI capabilities while introducing web search functionality that keeps conversations current with the latest information. Our native macOS AI chat application now offers more flexibility in AI model selection and real-time information access.
120 AI Chat now supports local models through LM Studio, expanding privacy-focused AI options for users who prefer complete control over their AI interactions.
LM Studio integration enables users to run sophisticated AI models entirely on their Mac hardware, ensuring sensitive conversations and proprietary information never leave their local environment.
We've added localhost server settings within the AI Providers Settings panel, giving users precise control over local AI connections and server configurations.
Custom URL configuration: Users can now specify exact local URLs for both Ollama and LM Studio connections, enabling advanced setups with custom ports, authentication, and network configurations that suit complex local AI deployments.
120 AI Chat now provides web search functionality that enables AI models to access current information during conversations, bridging the knowledge gap between AI training data and recent developments.
DuckDuckGo integration: We've chosen DuckDuckGo as our default search provider to maintain user privacy while providing comprehensive web search capabilities. The system automatically scrapes the first 3 search results to provide AI models with current information.
User-controlled activation: Web search can be enabled or disabled during a conversation, giving users complete control over when AI models should access external information versus relying solely on trained knowledge.
Context preservation: Search results integrate seamlessly with conversation flow, allowing AI models to reference current information while maintaining the natural conversation rhythm that users expect.
Privacy-focused approach: The DuckDuckGo integration ensures that web searches don't track user queries or create persistent search profiles, maintaining the privacy standards that 120 AI Chat users expect.
We've resolved rendering issues affecting file preview icons and file names within user messages, ensuring that attached PDFs and images display correctly throughout conversation history.
File names and type indicators display accurately, making it easier for users to reference and preview attached documents within lengthy conversations.
Development of 120 AI Chat for Windows continues with focus on performance optimization and feature parity with our macOS version. Our team is conducting extensive testing to ensure that Windows users receive the same high-quality, responsive experience.
We're working to optimize AI model switching, multimodal content handling, and conversation management for Windows-specific performance characteristics. All core features including multimodal support, local AI integration, and web search capabilities are being verified for Windows compatibility and optimal performance.
Auto-update packaging: The Windows version will launch with integrated auto-update functionality, eliminating the manual download and installation process that creates friction for regular users when new versions are released.
Next week, we're continuing intensive Windows development with focus on packaging the final release version. The Windows version will bring the complete 120 AI Chat experience to a broader user base while maintaining our commitment to native performance and platform-specific optimization.
We're also exploring advanced web search customization options that will allow users to configure search providers, result quantities, and integration methods to suit specific research workflows and privacy preferences.