Perplexica: Redefining Search with an Open-Source AI Engine

🧠 What is Perplexica?

Perplexica is an open-source AI-powered search engine developed by ItzCrazyKns, designed to provide a smart search experience without relying on closed platforms. By combining large language models (LLMs), embedding-based similarity search, and the SearxNG meta-search engine, it enables users to interact using natural language, fetch relevant web content in real time, and generate informative, source-backed responses.

⚙️ Key Features

🔄 Multi-Mode Search Experience

Normal Mode: Processes user queries and performs standard web searches.
Copilot Mode (under development): Enhances discovery by generating multiple search queries to retrieve more comprehensive information.

🎯 Six Specialized Search Modes

Global Search: Retrieves the most relevant results from the open web.
Writing Assistant: Assists in generating content without the need to search online.
Academic Search: Helps locate academic papers and literature for research.
YouTube Search: Finds and summarizes video content on YouTube.
Wolfram Alpha Queries: Handles math and data-driven queries via Wolfram Alpha.
Reddit Search: Retrieves related discussions and insights from Reddit.

🧩 Local LLM Support

Supports integration with local LLMs such as Llama3 or Mixtral using Ollama, enabling more private and responsive query processing.

🔗 API Integration

Developers can integrate Perplexica’s intelligent search capabilities into their own applications using its API.

🕒 Real-Time Web Updates

Leverages SearxNG to fetch the latest information, solving the problem of outdated data common in traditional search engines.

🧬 Technical Architecture

Here’s how Perplexica works under the hood:

User Interaction: Users send queries through WebSocket to the backend.
Query Handling: The system analyzes chat history and questions to decide whether to search the web and formulates the appropriate queries.
Web Search: SearxNG fetches relevant search results from the internet.
Similarity Search: Queries and retrieved content are embedded and matched to identify the most relevant resources.
Response Generation: The system combines history, queries, and content to generate a contextual response.
Display Output: The final answer is delivered through a user-friendly interface.

This modular architecture ensures Perplexica is both powerful and extensible.

🔗 Project Repository

GitHub: https://github.com/ItzCrazyKns/Perplexica

🌐 Use Cases

Private Personal Search: A self-hosted search engine that protects your privacy while offering smart results.
Academic Research: Useful for finding papers, references, and scholarly sources.
Enterprise Knowledge Management: Integrate Perplexica to help employees quickly find company-related documents.
Developer Integration: Add intelligent search features into your apps with the provided API.
Education & Learning: A great companion for students and teachers to find relevant resources quickly.