PipesHub AI is an open-source workplace AI platform designed to help enterprises unify and search their scattered work data across various applications. It provides natural language search, question answering, and insights, functioning as an alternative to proprietary solutions like Glean. The platform aims to boost employee and AI agent efficiency by centralizing information and offering a no-code interface for custom app and AI agent creation.

How It Works

PipesHub AI leverages a knowledge graph backbone to structure data from diverse sources, enabling efficient retrieval and providing cited results through page ranking. Its architecture is modular and scalable, allowing for independent service scaling. Users have the flexibility to integrate their preferred deep learning models for indexing and inference, supporting both real-time and scheduled data indexing. Access controls are maintained at the source level, ensuring data visibility is restricted to authorized personnel.

Quick Start & Requirements

• Installation: Deployment is managed via Docker Compose. Users can clone the repository and use docker-compose.dev.yml for development or docker-compose.prod.yml for production.
• Prerequisites: Docker and Docker Compose are required. Environment variables for secrets, passwords, and service URLs need to be configured.
• Setup: Detailed deployment guides are available for both development and production environments.

Highlighted Details

• Supports a wide range of enterprise connectors (Google Workspace, Microsoft 365, Slack, Jira, Confluence, etc.), with many planned for release.
• Handles various file formats, including PDFs (with OCR), Office documents, and text-based files, with image, audio, and video support coming soon.
• Offers enterprise-grade security with access-driven visibility and built-in data security.
• Provides deployment flexibility for both on-premise and cloud environments.

Maintenance & Community

The project encourages community contributions and provides a guide for setting up development environments and contribution workflows.

Licensing & Compatibility

The repository does not explicitly state a license in the provided README.

Limitations & Caveats

Support for several key connectors (e.g., OneDrive, SharePoint, Outlook, Slack, Notion, Jira, Confluence, MS Teams) is listed as "coming next month," indicating these features are not yet fully available. Image, audio, and video file format support is also marked as "Coming Soon."