A project brief, a folder of notes and several browser tabs can leave the final report harder to assemble than the research itself. Orkas brings those working materials into a local-first desktop AI workspace so a project can move from a question to a reviewable document or editable presentation through one coordinated conversation. For an illustrative workflow, give Orkas a product brief and source documents, then ask Commander to organize the work: research the unanswered questions, draft a report and prepare slides from the findings. DeepResearcher supports source-based research, while writing and presentation specialists develop the next deliverables. Review the sources, claims and exported files before sharing them; this example describes an intended workflow, not a measured customer result or an independent test. Orkas is open source under the MIT license. The core desktop app is free, with packaged builds for macOS and Windows. Users can inspect the public code and connect supported model providers using their own API keys, or choose Orkas-managed models. Model usage and optional cloud services may cost extra. Local project files keep inputs and outputs together, but requests to external model providers still follow those providers' data handling policies. This product information was prepared with AI assistance on behalf of the Orkas project. Visit the official site to check current features, downloads and provider requirements.