NexaSearch

An open-source vertical search engine for developers, researchers, and AI practitioners.

We don't compete in data. We compete in depth.

Demo

--------->COMING<---------

Why Nexasearch?

Google is excellent for general web search.

But developers often need to search:

• GitHub Issues • Pull Requests • Documentation • Research Papers • PDFs • API References

NexaSearch brings these sources together into one search experience.

Features

• Hybrid Search (BM25 + Vector)

• Research Paper Search

• GitHub Search

• Documentation Search

• PDF Search

• Live Crawling

• Semantic Retrieval

• Cross Encoder Reranking

• Open Source

Language & Technology Used

• Python
• CSS/JS
• Crawler
• RAG

Project Structure

NexaSearch/
├── .env.example
├── .gitignore
├── app_streamlit.py
├── cli.py
├── pyproject.toml
├── readme.md
├── uv.lock
├── cleaner/
│   └── distil_robert.py
├── data/
│   ├── index/
│   ├── nexasearch.db
│   └── raw/
│       ├── json/
│       ├── markdown/
│       ├── pdf/
│       └── sitemap/
├── doc/
├── evals/
│   └── evaluator.py
├── sources/
│   └── sources.py
├── src/
│   ├── Indexing/
│   │   ├── ranking.py
│   │   └── vector_index.py
│   ├── chunking/
│   │   └── chunker.py
│   ├── config/
│   │   └── settings.py
│   ├── crawler/
│   │   ├── crawler.py
│   │   ├── robots.py
│   │   └── sitemap.py
│   ├── embedding/
│   │   └── embedder.py
│   ├── pipeline/
│   │   └── orchestrator.py
│   ├── processing/
│   │   ├── cleaner.py
│   │   └── deduplication.py
│   ├── quality/
│   │   └── quality_filter.py
│   ├── search/
│   │   └── engine.py
│   ├── seeder/
│   │   └── url_seeder.py
│   └── storage/
│       └── duckdb_store.py
└── tests/
    ├── test_chunker.py
    ├── test_cleaner.py
    ├── test_deduplication.py
    ├── test_indexing.py
    ├── test_lsh.py
    └── test_quality.py

Workflow

Seed Sources
      │
      ▼
 Crawl Pipeline
      │
      ▼
 Data Processing
      │
      ▼
 Hybrid Index
      │
      ▼
 Search
      │
      ▼
   Top 50
      |
 AI Answer (Optional)

-> For the complete system design follow docs/architecture.md

Installation

1. Fork the Repository

Fork this repository to your GitHub account by clicking the Fork button.

2. Clone the Repository

Clone your fork to your local machine.

git clone https://github.com/<niteshver>/NexaSearch.git

3. Navigate to the Project Directory

cd NexaSearch

4. Create a Virtual Environment

Linux / macOS

python3 -m venv .venv
source .venv/bin/activate

Windows

python -m venv .venv
.venv\Scripts\activate

5. Install Dependencies

using uv:

uv sync

6. Configure Environment Variables

Create a .env file in the project root and add the required configuration.

# Example
OPENAI_API_KEY=your_api_key

7. Run the Project

example