Onyx Project Detailed Introduction

Project Overview

Onyx (formerly Danswer) is an open-source generative AI enterprise search platform that provides teams with an AI platform to connect company documents, applications, and people. It offers a feature-rich chat interface that can connect to any Large Language Model (LLM) of your choice.

Simply put, Onyx is like a ChatGPT with access to your team's unique knowledge.

Core Features

1. Intelligent Search & Q&A

• Deep Research Capabilities: Ability to conduct in-depth research and analysis of team knowledge.
• Secure AI Chat: Provides a secure AI chat interface, supporting any LLM.
• Personalized Search: Offers a personalized search experience based on user needs.

2. Broad Data Source Connectivity

Supports 40+ connectors, keeping knowledge and access control in sync, including:

Office Collaboration Tools

• Google Drive
• Microsoft Sharepoint
• Dropbox
• Local Files

Communication Platforms

• Slack
• Microsoft Teams
• Gmail

Project Management

• Confluence
• Jira
• Github

Customer Relationship Management

• Salesforce
• Zendesk
• Gong

Other

• Websites
• More connectors...

3. Custom AI Agents

Create custom AI agents with unique prompts, knowledge, and operational capabilities.

4. Flexible Deployment Options

Onyx can be securely deployed anywhere and at any scale - on a laptop, on-premises, or in the cloud.

Deployment Methods:

• On-Premises: Run with a single docker compose command.
• Cloud Deployment: Supports Kubernetes for high availability/scalable deployment.
• Laptop: Can even be run locally on a laptop.

Advanced Features

Security Features

• SSO Support: OIDC/SAML/OAuth2
• RBAC: Role-Based Access Control
• Credential Encryption: Ensures data security.

Knowledge Management

• Document Set Management: Supports the creation and management of document sets.
• Query History: Records and analyzes query history.
• Usage Analytics: Provides detailed usage analysis.

Technical Innovations

• Custom Deep Learning Models: Used for indexing and inference time.
• New Information Retrieval Methods: StructRAG, LightGraphRAG, etc.
• Code Search: Supports code repository search.
• SQL and Structured Query Language: Supports structured data queries.

Product Editions

Community Edition (CE)

• License: MIT Expat License
• Features: Basic features are free to use.
• Deployment: Follow the deployment guide.

Enterprise Edition (EE)

• Target Users: Large organizations
• Additional Features: Provides enterprise-grade features.

Technical Architecture

Supported Scale

• User Scale: Tested to support tens of thousands of users.
• Document Scale: Supports hundreds of millions of documents.
• High Availability: Achieved through Docker or Kubernetes.

Core Technologies

• Retrieval Augmented Generation (RAG): Combines search and generation capabilities.
• Vector Search: Efficient semantic search.
• Multi-Modal Support: Supports multiple data types such as text and images.

Use Cases

Enterprise Knowledge Management

• Quickly find company documents and information.
• Cross-departmental knowledge sharing.
• New employee training and knowledge transfer.

Team Collaboration

• Real-time access to project-related information.
• Automatically summarize meeting notes and documents.
• Intelligent recommendation of relevant resources.

R&D Support

• Code repository search and analysis.
• Technical documentation query.
• Expert identification and advice.

Quick Start

Cloud Trial

Visit Onyx Cloud for a free trial.

Local Deployment

# One-click deployment using Docker Compose
docker compose up

Detailed Documentation

View the Official Documentation for complete deployment and usage guides.

Summary

Onyx is an open-source AI enterprise search platform that helps teams find information across all company data. Through powerful AI capabilities and broad data source connectivity, it provides enterprises with an intelligent knowledge management and search solution, truly realizing the vision of "a ChatGPT with your team's unique knowledge."