holaOS

Overview of holaOS

What it is

holaOS is an open-source AI agent workspace. It allows users to run agents like Claude Code and Codex alongside their applications, tools (with over 100 integrations plus MCP support), browser, and files. The workspace features shared memory and is local-first. Users can use built-in models or bring their own API keys.

Key features

• HolaApps: install apps from an in-workspace marketplace that open as live, interactive surfaces beside the agent.
• IM integration: connect chat tools like Slack, Feishu, DingTalk, and WeChat so the agent reads real conversations.
• Skills, integrations, and MCP: connect services like Gmail, Notion, and GitHub with one-click OAuth, plug in any MCP server, and package workflows as skills.
• Model flexibility: built-in frontier models (Kimi K3, GLM 5.2, GPT 5.6, Claude Opus 5, Fable 5) or bring your own keys for OpenAI, Anthropic, or compatible endpoints.
• Multi-agent support: run Claude Code, Codex, or the built-in holaOS agent side by side with shared memory, tools, and skills.
• Local-first memory: context, preferences, and project history are stored locally as plain editable files.
• Agent-operable workstation: includes a real browser, frontier media generation, real deliverable file formats (.xlsx, .pptx, .docx), and scheduled automation.

Use cases

holaOS can be used for workflows where an AI agent needs to interact with multiple applications, chat platforms, and custom integrations. It is suitable for tasks such as automating data extraction from a browser, generating reports as real documents, creating media from prompts, and scheduling recurring work. The local-first nature makes it appropriate for teams or individuals who want to keep their data on their own machines.

Getting started

See the repository README.

License and pricing

The license is not stated. The software is marked as open source.

Based on the project's public README and repository metadata.