OpenPerPlex is an open-source AI search engine designed for web search, targeting users who want advanced semantic search capabilities. It aims to provide more relevant results through AI-powered features like semantic chunking and result reranking.

How It Works

The backend utilizes semantic chunking via the semantic-chunkers library and Cohere for processing search queries. Results are reranked using the JINA API, and Google search integration is handled by serper.dev. Inference is powered by Groq, leveraging the Llama 3 70B model for advanced language understanding and response generation.

Quick Start & Requirements

• Install: pip install -r requirements.txt
• Prerequisites: Python 3.11+, API keys for Cohere, JINA, Serper.dev, and Groq.
• Running: uvicorn main:app --port 8000
• Documentation: https://github.com/YassKhazzan/openperplex_backend_os

Highlighted Details

• Semantic chunking with Cohere and semantic-chunkers.
• Result reranking via JINA API.
• Google search integration through serper.dev.
• Inference engine: Groq with Llama 3 70B model.

Maintenance & Community

Contributions are welcomed via issues and pull requests. Contact is available through GitHub issues.

Licensing & Compatibility

Licensed under the MIT License, permitting commercial use and integration with closed-source projects.

Limitations & Caveats

Requires multiple third-party API keys, which may incur costs. The project relies on external services for core functionality, making it dependent on their availability and pricing.