holaboss-ai/holaOS

Open-source agentic workspace enterprises can make their own. Connect the systems you already run — 100+ integrations, MCP, chat tools, apps, browser, local files — with shared memory. Any agent (Claude Code, Codex), any model, or BYOK. Set up in clicks, not months. Local-first: your data never leaves your machines.

What it solves

HolaOS is an agentic workspace designed to integrate AI agents directly into a user's existing professional workflow. It solves the problem of "chat-only" AI interfaces by providing a side-by-side environment where agents can operate real applications (like Notion or browsers) and access real-time communication tools (like Slack or WeChat) without the user having to manually copy-paste context between different windows.

How it works

The system operates as a local-first desktop application (built with Electron and TypeScript) that acts as an orchestration layer for various AI components:

  • HolaApps: Interactive UI surfaces that allow agents to drive real apps side-by-side with the chat interface.
  • IM Integration: Direct connections to chat tools (Slack, Feishu, DingTalk, WeChat) so agents can read conversation context.
  • Extensibility: Uses the Model Context Protocol (MCP) and a custom "Skills" system to add new capabilities and workflows.
  • Shared Memory: A local-first memory system stored as plain files, allowing multiple agents (e.g., Claude Code, Codex, or the built-in agent) to share the same project history and preferences.
  • Model Flexibility: Supports built-in frontier models or "Bring Your Own Key" (BYOK) for providers like OpenAI and Anthropic.

Who it’s for

It is for teams and individuals who want a highly customizable AI workstation where agents have actual operational power over their tools and data, while maintaining strict local control over their privacy and memory.

Highlights

  • Side-by-Side UI: Agents operate real app surfaces rather than just providing text transcripts.
  • Local-First Privacy: Workspace files, memory, and embeddings are stored on the user's disk, not in the cloud.
  • Agent Agnostic: Allows running different agents (Claude Code, Codex, etc.) within the same workspace using shared tools and memory.
  • HolaHub: A recipe platform for sharing and installing pre-configured agent setups (models, prompts, and skills) in one click.
  • Direct Deliverables: Capable of generating real files like .xlsx, .pptx, and .docx instead of just markdown text.

Related

  • Project
    jaylfc/taOS

    A self-hosted AI agent platform that turns spare hardware into a distributed compute cluster, providing a framework-agnostic memory system and a full web-based desktop interface.

  • Project
    liliMozi/openhanako

    HanaAgent is a user-friendly AI assistant with memory and personality that can operate your computer, browse the web, and collaborate across multiple agents via a graphical interface.

  • Project
    shiwenwen/hope-agent

    A local-first personal AI agent that autonomously executes long-term goals, manages a persistent knowledge base, and controls desktop/browser interfaces.

  • Project
    openyak/openyak

    A universal interface for AI agents that allows users to maintain a single persistent chat while switching between different agent runtimes like Claude Code and Codex.

  • Project
    openagents-org/openagents

    An open-source collaborative OS for AI agents that provides a unified workspace where multiple agents can share threads, files, and a browser to work together.