Orkas

What it does

Orkas is a desktop-first, open-source workspace for coordinating specialist AI agents to generate and edit images, video, slides and documents from briefs, references or supplied files. It’s aimed at creative teams and producers who want agent-driven automation while keeping files and compute under their control. Multi-agent orchestration Orkas runs a Commander agent that dispatches specialist subagents (research, writing, video, image, code, slides and data). Describe the task once and the Commander coordinates agents to execute parts of the workflow, chain steps and return consolidated outputs for review. VideoStudio and ImageStudio Built-in VideoStudio supports creating and editing animations, captions, narration and reviewable cuts from raw footage; ImageStudio produces supporting visuals and variants. The UI and agent workflows are tuned to assemble and iterate creative assets rather than only producing single-shot renders. Skills, automations and connectors Orkas exposes reusable skills and connectors for APIs, file readers, OCR, CLIs and third-party model providers. Users can install and reuse local packages for common steps, automate repetitive transforms (batch decks, multiple ad variants) and extend the system with custom skills. Local-first desktop model options The core app runs on desktop platforms and emphasizes local file access and environment setup. You can configure your own model provider API keys (BYO) or use Orkas-managed models; some integrations and advanced tools are provided as local packages installed on demand. This design favors control and local workflows over a pure cloud drop-in. Product outputs and productivity use cases Orkas can generate text, images, video cuts, PPTX slide decks, analyzed spreadsheets and reviewable creative packages. Typical jobs include drafting briefs from research, converting outlines into slide decks, producing localized creative variations, and assembling short review videos from supplied footage.