wtf is staffdeck?

StaffDeck is an enterprise platform for building what it calls digital employees, meaning software agents built from a person's actual work experience, business processes, and decision criteria. The idea is to turn one employee's know-how into a reusable, traceable asset that the organization keeps even if that person moves on, rather than a one-off script or a personal productivity trick. It was developed jointly by ModelBest, a data intelligence lab at Northeastern University, Tsinghua's NLP lab, OpenBMB, and AI9Stars. Each digital employee gets a position, an employee ID, a capability profile, and a work history, and can grow its capabilities over time while permissions keep it isolated from things it should not touch. Complex work is handled through state machine driven procedures generated from plain language instructions, which can switch between multiple flows, keep context, and be edited visually with version history. For finding information, StaffDeck builds a navigable index across documents at multiple levels, chapters, pages, and summaries, so a digital employee can narrow down where an answer likely lives before pulling the exact source text, and every answer can cite where it came from. Digital employees can act on real systems through HTTP APIs, MCP, and scheduled tasks, and the platform keeps long-term memory, full execution traces, a way for a human to take over, and feedback analysis to improve over time. StaffDeck ships as a desktop app for macOS, Windows, and Linux, or it can be run from source with Python 3.11 or newer and Node.js 20 or newer. Getting it running locally means cloning the repository, installing backend and frontend dependencies, configuring an OpenAI compatible model endpoint and API key in an environment file, and starting a single FastAPI process that serves the web interface, API, and documentation together. The project is licensed under the GNU AGPL-3.0 license.