I Built an Open-Source AI Workspace Operating System in a Few Weekends

I Built an Open-Source AI Workspace Operating System in a Few Weekends

Over the last few weekends, I built LOOP.

It started from a simple frustration.

My workflow was spread across multiple tools:

• Slack for communication
• Notion for documentation
• Jira for project management
• GitHub for development
• ChatGPT for AI assistance
• Google Workspace for operations

Every tool had its own context, data, and workflow.

The AI had no understanding of what was happening across the workspace.

So I started asking:

What if AI wasn't another tab?

What if AI was part of the operating system itself?

That idea became LOOP.

What is LOOP?

LOOP is an open-source AI-native workspace operating system.

It combines:

• Projects and task management
• Team chat and channels
• Collaborative documents
• Multi-agent AI teams
• Workspace memory
• MCP-powered tools
• Human approval workflows
• Google Workspace integrations
• GitHub integrations
• Finance and invoice management
• Workspace intelligence and health monitoring

The Architecture

LOOP is built with:

• FastAPI
• Next.js
• PostgreSQL
• Redis
• Docker
• MCP (Model Context Protocol)
• OpenRouter

At the core is Hermes, an AI agent system capable of:

• Managing workspace memory
• Coordinating specialized agents
• Using MCP tools
• Creating approval-based actions
• Interacting with external systems

Why Open Source?

Honestly, I don't know if anyone else needs this.

I originally built it for internal use.

Instead of spending months guessing what users want, I decided to open-source it and let the community decide whether the idea has value.

Looking for Feedback

I'm interested in hearing:

• Would you use something like this?
• Which features are most interesting?
• What's missing?
• What would prevent adoption?

GitHub:repo

Feedback, criticism, and ideas are all welcome.