xynehq/xyne

What it solves

Xyne addresses the problem of fragmented work information spread across various SaaS applications, documents, files, repositories, and communication tools. It provides a centralized, AI-powered search and answer engine that allows users to find information and get up-to-date answers with sources across their entire organization's data.

How it works

Xyne connects to various work applications (such as Google Workspace, Slack, and the Atlassian suite) to securely index data and map relationships between them. It acts as a context provider for LLMs, enabling a search-and-answer experience similar to a combination of Google and ChatGPT. The system is model-agnostic, meaning it can be integrated with any LLM or cloud provider, including local models via Ollama.

Who it’s for

It is designed for organizations and teams that need a secure, private, and permissions-aware way to integrate AI into their internal workflows and data management.

Highlights

• Self-hosted: Can be deployed on laptops, on-premises, or in the cloud.
• Model Agnostic: Compatible with any LLM of choice, including local models like DeepSeek via Ollama.
• Permissions-aware: Enforces existing application permissions so users only see data they have access to.
• Private & secure: No telemetry and no training on user data or prompts.
• Google Workspace Integration: Full integration across Drive, PDFs, docs, sheets, slides, calendar, contacts, and Gmail.
• High performance: Uses multi-threaded data ingestion.