Home
Login

AI-powered search and RAG platform that doesn't require data movement, providing instant answers to enterprise knowledge securely across 100+ applications.

Apache-2.0Python 2.8kswirlai Last Updated: 2025-06-19

Swirl-Search Project Details

Project Overview

Swirl-Search is an open-source AI search and RAG (Retrieval-Augmented Generation) platform designed to provide enterprises with a unified search experience across 100+ applications, securely obtaining instant answers to enterprise knowledge without moving data. Built on Python and Django, the project can be deployed in minutes.

Core Features

1. Unified Search Experience

  • Multi-Source Integration: Unified search across databases (SQL and NoSQL), cloud services, search providers, data silos, and tools like Miro, Jira, GitHub, etc.
  • No Data Migration: Data remains in its original location, without copying or moving.
  • Permission Protection: Search results adhere to existing permission systems.

2. AI-Powered Intelligent Search

  • RAG Technology: Uses configured LLM embedding techniques to re-rank results from all response sources.
  • Intelligent Q&A: Provides instant answers with source links.
  • Contextual Understanding: Understands enterprise-specific context to provide accurate answers.

3. Enterprise-Grade Security and Deployment

  • Data Security: Sensitive data remains secure, without external storage.
  • Rapid Deployment: Deploys in minutes, not months.
  • Infrastructure Control: Runs on your infrastructure, providing complete control.

Technical Architecture

Core Components

  • Search Providers: Supports 100+ enterprise connectors.
  • Query Processing: Intelligent query transformation and adaptation.
  • Result Aggregation: Asynchronous search federation and result re-ranking.
  • AI Enhancement: Integrates OpenAI, Hugging Face, and other AI services.

Supported Data Sources

  • Office Suites: SharePoint, Confluence, Google Drive.
  • Development Tools: GitHub, Jira, Documentation Systems.
  • Databases: SQL Databases, NoSQL Databases.
  • Cloud Services: Various cloud platforms and APIs.
  • Search Engines: Apache Solr, Elasticsearch, etc.

Installation and Deployment

Docker Quick Deployment

# Download the configuration file
curl https://raw.githubusercontent.com/swirlai/swirl-search/main/docker-compose.yaml -o docker-compose.yaml

# Start the service
docker-compose pull && docker-compose up

Environment Requirements

  • Docker application (latest version).
  • Windows users require WSL 2 or Hyper-V backend.
  • Optional: OpenAI API key for RAG functionality.

Use Cases

1. Enterprise Knowledge Management

  • Connect SharePoint, Confluence, and Drive.
  • Get instant answers with source links.
  • Keep sensitive data secure.

2. Customer Support

  • Search support documentation and tickets.
  • Draft replies using enterprise content.
  • Maintain consistent answer standards.

3. Development Teams

  • Search GitHub, Jira, and documentation.
  • Find code examples and solutions.
  • Accelerate development workflows.

4. Unified Search Portal

  • Unified search across all tools.
  • Results adhere to existing permissions.
  • No data duplication required.

Key Advantages

Performance Advantages

  • Time Savings: Teams using SWIRL save an average of 7.5 hours of production time per week.
  • Fast Response: Returns ranked results in seconds.
  • Asynchronous Processing: Supports synchronous and asynchronous search federation.

Technical Advantages

  • Open Source: Fully open source, freely customizable.
  • Modular: Supports the extension of processors, connectors, and mixers.
  • Intelligent Processing: Includes spell correction, duplicate detection, relevance ranking, and more.

Enterprise Advantages

  • Security: Data does not leave the enterprise environment.
  • Scalability: Supports large-scale enterprise deployments.
  • Flexibility: Supports multiple data sources and custom configurations.

Technical Features

Intelligent Query Processing

  • Query transformation and rewriting.
  • Stemming and stop word processing.
  • Spell correction support.

Result Optimization

  • Duplicate detection based on cosine similarity.
  • Multiple sorting strategies (relevance, date, polling).
  • Real-time result filtering.

Extensibility

  • Custom processor development.
  • Connector plugin system.
  • Configurable result mixers.

Summary

Swirl-Search is a powerful enterprise-grade AI search solution that helps enterprises quickly access knowledge scattered across various systems through unified search, intelligent Q&A, and secure deployment. Its open-source nature, rapid deployment capabilities, and powerful AI features make it an ideal choice for modern enterprise knowledge management.