research / agent

Open Glean

A web app for asking questions across your Hydra DB memories, files and connected apps, with answers that show their sources.

Repository owner hydra-db

Not rated yet
0 comments

Like 0 Save

Published 11 Oct 2026Listing updated 12 Oct 2026

At a glance

Purpose: A web app for asking questions across your Hydra DB memories, files and connected apps, with answers that show their sources.
Runs in: Web service
Operating systems: Not stated
Needs: Runtime: Not stated; Needs internet: Yes; Needs accounts or API keys beyond your agent: Yes
Files and data: Reads files: Yes; Writes files: Not stated; Sends data to other services: Yes
Tool type: Agent
Delivery: From GitHub
Version: Latest on the main branch
Access: Code on GitHub
Compatibility: The README identifies a Next.js 16 and React 19 app with Tailwind CSS v4 and @hydradb/sdk. The README describes support for OpenAI-compatible endpoints and a built-in OpenRouter model picker. The README describes MongoDB chat persistence and says Amazon DocumentDB requires a standard MongoDB connection string or the Lambda proxy. The README says OPEN_GLEAN_ALLOW_PRIVATE_LLM_URL enables private or loopback LLM endpoints, such as Ollama or LM Studio, and is off by default. The README provides a Docker route and names no operating systems.
External costs: According to the README, you need a Hydra DB API key and, for answers, an OpenRouter key or another OpenAI-compatible key. It also names an optional MongoDB connection for chat history. The README states no prices. It says each Deep Research run costs many LLM and retrieval calls. The services you connect may charge for usage.
Permissions: The README's local setup installs dependencies with npm ci and asks for a Hydra DB API key and a model-provider key. The README also provides a Docker build and run route with deployment-level keys. The README says queries read the selected Hydra database and collections, and describes adding files, notes and URLs and starting connector syncs. The README's request diagram sends Hydra requests through the server to Hydra's API and chat requests to the chosen LLM provider. The README describes optional web search through OpenRouter's web plugin and chat persistence in MongoDB with a localStorage fallback. The README says private or loopback LLM endpoints are disabled by default.
Usage terms: Apache License 2.0, as stated in the repository. See its LICENSE file for the exact terms. The README adds that connector logos are third-party trademarks that the Apache grant does not cover, and it points to the NOTICE file.
Interface: web

Code on GitHub

hydra-db/open-glean

Apache-2.0

TypeScript1,587 stars496 forksLast pushed 6 Oct 2026 UTC

An open-source AI platform for knowledge work. Connect your apps, find answers, and get work done.

Public GitHub data fetched 13 Oct 2026 UTC · 0 days old

View on GitHub ↗

What it does

According to the README, Open Glean is an AI workspace over Hydra DB for questions across memories, files and connected apps. The README says it retrieves context, writes answers and cites sources.

The README describes Ask as streaming answers with inline citations, a sources panel and optional web search through OpenRouter's web plugin. The README says Deep Research plans sub-questions, runs retrieval in parallel, writes branch findings and produces an answer. The README also describes database and collection selection, a Context page, a mindmap and Hydra connector integrations.

The README identifies the stack as Next.js 16, React 19, Tailwind CSS v4 and @hydradb/sdk, and provides local and Docker setup routes.

Who it is for

People who use Hydra DB and want a web workspace to ask questions across their data, and developers who want to run or change the code.

Example tasks

Ask a question across your memories, files and connected apps and read the answer with its citations.
Run Deep Research on a question that one query cannot answer.
Pick the database and collections that a query searches from the top bar.
Add files, notes or URLs from the Context page.
Connect one of Hydra's connectors in the app and start syncing.

Getting started

The README gives a local route and a Docker route. For the local route, run npm ci and then npm run dev. Copy .env.example to .env.local and set at least OPEN_GLEAN_SESSION_SECRET. Open the app in a browser on port 3000 of your machine, paste a Hydra DB API key, pick a database, and add a model in Settings. The Docker route runs docker build -t open-glean . and then a docker run command that passes environment variables such as the session secret, the Hydra key and an OpenRouter key. The README has the exact commands.

Limitations

According to the README, a Hydra DB API key is needed, entered in the app or supplied through HYDRA_API_KEY. The README says search works without a model, but answers require OPEN_GLEAN_LLM_MODEL or a per-user model selection. The README marks OPEN_GLEAN_SESSION_SECRET as required in production. The README says conversations persist to MongoDB when reachable, with a localStorage fallback. The README says each Deep Research run costs many LLM and retrieval calls, with three concurrent runs per instance by default. The README warns that the IAM token path cannot authenticate against Amazon DocumentDB. The README says connector logos are third-party trademarks outside the Apache grant. The repository can change at any time; check its README before you install.

Data practices

The README's request diagram sends browser requests through the Next.js proxy and @hydradb/sdk to Hydra's API, and chat requests to the chosen LLM provider. The README says connected keys are held in an AES-256-GCM-encrypted, httpOnly session cookie, resolved server-side and pinned to their stored endpoint. The README says conversations persist to MongoDB when reachable, with a localStorage fallback, and non-secret settings stay in the browser. The README says web search uses the LLM provider's web plugin, with citations alongside Hydra sources. The README does not mention telemetry.