Studi’OS, coordination between humans and AI agents

When several people and several AI agents work on the same project, context gets lost: who does what, which decisions were made, what has already been tried. Studi’OS gives the whole group a shared, durable memory.

Make humans and agents cooperate from different workstations, without scanning the whole repository for every task and without two agents stepping on the same files.

An application core exposing an HTTP API and an MCP server for agents, backed by PostgreSQL, with S3 object storage for files sent directly through presigned URLs.

• Projects, tasks, sessions, resource claims and traced decisions.
• Shared project memory that agents can query.
• Python client with a local daemon, offline queue and command line.
• Web dashboard and desktop application (Tauri) supervising the daemon.