FreeAskInternet

Why we included this project

This project runs a Perplexity-style search-and-answer pipeline entirely on your own machine. You ask a question in a lightweight web interface, SearXNG gathers results across several search engines, and the crawled page text goes to an LLM that streams back an answer with references. There is no GPU requirement and no need for API keys out of the box: it can call free endpoints for ChatGPT, Qwen, Kimi, and GLM, or you can swap in a local Ollama or llama.cpp model instead. That makes it an easy way to get a private, self-hosted chat-search setup running on a laptop or Docker host, and a useful base for understanding how retrieval-augmented generation works end to end.