Windows: likely yes, but not yet. The current official positioning of Cowork is that it’s a research preview available through Claude Desktop on macOS, and it is not available on web or mobile right now. Public statements about the preview indicate plans to expand platform support, and Windows is explicitly called out as a target. Practically, you should treat today’s situation as “macOS-only,” and treat Windows availability as “planned,” meaning it could arrive later but with timing, feature parity, and rollout details that may change.
Linux and mobile: there is no stable guarantee today. The safest, technically accurate planning assumption is that Cowork’s deeper “agent runtime” (desktop integration, controlled file access, long-running tasks, and UI checkpoints) is more complex to ship cross-platform than standard chat, and vendors often expand one platform at a time. That means Linux and mobile support may come later—or may arrive in a different form than desktop Cowork. Also, the features that make Cowork valuable (folder access, file writes, running tasks that generate artifacts) map naturally to desktop environments; on mobile, those same capabilities can be constrained by OS sandboxing and user experience. So if you need Cowork-like behavior on non-macOS today, design your workflow so it doesn’t depend on the Cowork UI being present everywhere.
The most robust strategy is to make your larger system platform-neutral and treat Cowork as an optional accelerator. For example, keep your authoritative data flow as: export/snapshot → preprocess → validate → ingest → serve. Cowork can help with preprocess (cleaning docs, extracting fields, generating chunked outputs), but your ingestion and serving should not require Cowork to exist on every device. If your end product is semantic search or Q&A, embed and index the cleaned artifacts in a vector database such as Milvus or Zilliz Cloud. That way, when Cowork expands to Windows (and potentially other platforms), you can add it as a productivity layer without rewriting your core pipelines or changing how your system behaves in production.