hydra-db/open-glean

What it solves

Open Glean is an AI-powered workspace designed to act as a "second brain," allowing users to query their personal memories, files, and connected application data. It solves the problem of fragmented knowledge by providing a single interface to retrieve context and generate answers with citations from a variety of a user's own data sources.

How it works

The application is built as a Next.js server that acts as a proxy for the Hydra DB SDK. It retrieves context from Hydra DB and sends it to a user-provided OpenAI-compatible LLM (such as via OpenRouter) to generate answers. It supports a "Deep Research" mode that plans a Directed Acyclic Graph (DAG) of sub-questions to handle complex queries, running retrieval and generation in parallel across multiple branches.

Who it’s for

It is intended for individuals or teams who use Hydra DB to store their knowledge base and want a self-hosted, open-source AI interface to interact with their data using RAG (Retrieval-Augmented Generation).

Highlights

• Deep Research: Automatically breaks complex questions into a DAG of sub-questions for comprehensive answers.
• Mindmap: Visualizes the knowledge graph built by Hydra.
• Scope Switching: Allows users to target specific databases or collections for queries.
• Citations: Provides inline citations and a sources panel for all generated answers.
• BYO Model: Supports any OpenAI-compatible endpoint, including local models via Ollama or LM Studio.
• Integration Management: Manage Hydra connectors and sync data directly within the app.