FreeAskInternet_云技派,专注ai人工智能! 208

FreeAskInternet is a fully free, private, and locally running search aggregator that uses LLM to generate answers, with no GPU required. Users can ask questions, and the system will use searxng to perform multi-engine searches, then merge the search results into ChatGPT3.5 LLM to generate answers based on the search results. All processes run locally, requiring no GPU or Openai or Google API keys.

FreeAskInternet Features:

• Completely free (no API keys required)

• Fully localized (no GPU needed, any computer can run it)

• Completely private (everything runs locally, using a custom llm)

• No LLM hardware required to run (no GPU needed!)

• Uses the free ChatGPT3.5 / Qwen / Kimi / ShipuAI(GLM) API (no API keys needed! Thanks to OpenAI)

• Support for custom LLM (ollama, llama.cpp), yes, we love ollama!

• Quick and easy deployment with Docker Compose

• Web and mobile-friendly interface, designed for AI chat with web search augmentation, allowing easy access from any device.

How does FreeAskInternet work?

1. The system receives the user's input question through the FreeAskInternet UI interface (running locally), and calls searxng (running locally) to perform searches on multiple search engines.

2. It scrapes the content from the search result links and passes it to ChatGPT3.5/Kimi/Qwen/ZhipuAI/ollama (using a custom llm), asking the LLM to answer the user's question based on this content as reference.

3. The answer is transmitted to the chat UI.

4. We support custom LLM settings, so theoretically unlimited llm support.