llama-server configured for SearXNG live search
The author explains how they configured a local LLM to perform live web searches for free without subscriptions, eliminating the need for cloud-based services like ChatGPT and Gemini for real-time information queries. By running a SearXNG container, an MCP bridge, and enabling tool-calling flags in llama-server, the local model automatically detects when a query requires current data and retrieves it independently. Open article →