Perplexica: AI-Powered Open-Source Search Engine
Perplexica is an innovative open-source search engine powered by AI and machine learning. Inspired by Perplexity AI, it goes beyond traditional web searches by understanding user queries and providing clear, well-sourced answers. Unlike conventional search tools, Perplexica leverages advanced technologies like similarity searching and embeddings to refine results and ensure relevance.
Key Features
- Local LLM Support: Integrates with local models like Llama3 and Mixtral via Ollama
- Two Search Modes:
- Copilot Mode (in development): Enhances search by generating multiple queries for comprehensive results
- Normal Mode: Traditional web search processing
- Specialized Focus Modes: Six distinct modes including Academic Search, YouTube Search, and Wolfram Alpha integration
- Privacy-Focused: Uses SearxNG metasearch engine to ensure up-to-date results without compromising user privacy
- API Access: Developers can integrate Perplexica's capabilities into their applications
Technical Highlights
Perplexica employs a sophisticated architecture that combines:
- Next.js for the frontend
- Machine learning for result refinement
- SearxNG for real-time, privacy-respecting web searches
The project offers flexible deployment options, with Docker being the recommended method for simplified setup. Perplexica continues to evolve with planned features like history saving, additional focus modes, and enhanced Copilot functionality.
As an open-source project, Perplexica welcomes contributions and feedback from the community, embodying the principle that powerful AI tools should be accessible to everyone.