Nakama is an open-source AI agent platform for teams, hosted on GitHub with about 480 stars under the MIT license and active development in 2026. (It is an AI workspace, unrelated to the game server that shares the name.) Teams chat with agents in a dashboard that feels like ChatGPT, while the platform runs the agents behind it.
Each agent in Nakama can use tools, skills, and memory, and can be reached through channels such as Telegram, WhatsApp, Discord, and Slack, with its own identity and permissions. It was designed for teams from the start: one server can host many organizations, with invites and roles, and agents are isolated from each other so one team’s agent cannot see another’s data. Deployment is deliberately simple, either as a single Docker container you self-host or through the managed cloud offering, and the project positions itself as a multi-user alternative to single-operator agent setups.
The project explicitly courts teams that have outgrown single-operator agent frameworks: instead of one person’s agent on one laptop, a company runs one Nakama server and gives every team its own agents, channels, and permissions. Agents keep their own identities, so a sales agent and an engineering agent can coexist on the same server without sharing memory or tools they should not have.
Nakama is free and open source under the MIT license for self-hosting. A managed cloud hosting option is available for teams that would rather not run the server themselves.
It suits companies and teams that want shared AI agents the whole organization can build on and talk to through familiar chat channels, with proper separation between teams instead of one personal agent on one machine.