holaOS is an open-source agentic workspace that enterprises can make their own: a local-first environment where AI agents work across the systems a company already runs. It connects more than 100 integrations, chat tools, apps, the browser, and local files into one workspace with shared memory, so agents keep context instead of starting from zero every session.
The workspace is agent- and model-agnostic: bring Claude Code, Codex, or other agents, connect any major model or use your own keys, and let agents collaborate over shared memory and tools through MCP. Because it is local-first, company data stays on your own machines rather than being shipped to a third-party cloud. Setup is designed to take clicks rather than months of integration work, which makes it attractive to service companies that want “AI employees” handling hand-offs between sales, delivery, and customer management. The project has more than 11,400 GitHub stars and is actively developed, with updates in August 2026. A hosted product, holaboss, is built on the same stack.
The holaOS runtime is open source. The hosted holaboss product starts free and scales by token-based credits, with enterprise plans for custom and on-premise deployments.
Service companies and operations teams that want agents working across their existing tools with shared memory, without surrendering data to a hosted platform.