SearXNG: A free internet metasearch engine
SearXNG is a free, privacy-respecting metasearch engine that aggregates results without tracking users. Hacker News embraces it for its open-source nature, self-hosting capabilities, and its growing utility in privacy-conscious AI agent setups. It offers a powerful alternative to mainstream search, albeit with some trade-offs in speed and direct result quality.
The Lowdown
SearXNG is an open-source, free internet metasearch engine designed to aggregate results from various search services and databases without tracking or profiling users. The project's GitHub page showcases its active development, detailing recent commits, file structure, and development policies, reflecting a community-driven effort to maintain and improve its privacy-focused search capabilities.
- Privacy-Focused Aggregation: It pulls search results from multiple sources, shielding users from direct profiling by those sources.
- Self-Hostable: Users can deploy their own instances, gaining full control over their search experience and data.
- Active Development: The repository shows frequent updates, including dependency bumps, bug fixes, and feature enhancements.
- Community Contributions: Evidenced by pull requests and issue discussions, the project benefits from collective input.
- Developer-Centric Information: The GitHub page primarily serves as a development hub, focusing on code, commits, and project management rather than a user-facing product description.
In essence, SearXNG positions itself as a robust option for those prioritizing digital privacy and control over their search data, offering a customizable and transparent alternative to corporate search engines.
The Gossip
Privacy Prowess & Personal Preferences
Many users laud SearXNG for its privacy features, often self-hosting to avoid Google and other trackers. They accept occasional slower performance or the need to solve CAPTCHAs as a worthwhile trade-off for the privacy and control it offers. Some even use it as their daily driver, customizing backends with tools like YaCY.
AI Integration & Agent Utility
A significant discussion revolves around using SearXNG to provide search capabilities for local AI models and agents. Users describe setups involving JSON APIs, custom wrappers like TinySearch or searxng-mcp, and integrating with llama.cpp for tool calling, highlighting its value in creating privacy-preserving RAG applications.
Metasearch Mechanics & Limitations
Commenters discuss the inherent challenges of metasearch, including engines occasionally failing or being slow. The original creator notes the limitations of the metasearch concept, pointing towards his new project, Hister. There's also debate about whether sending queries to multiple sources still constitutes privacy, though the consensus is that the user is less directly profiled.
Hister's Hopeful Horizons
The original creator of SearX (asciimoo) introduced his new project, Hister, a personal full-text indexer for websites and local files. This sparked significant interest, with users seeing it as a solution for personal knowledge management, local "memory," and a potential future for decentralized general internet search, acknowledging its current limitations with data volume and file types.