Namzu vs CrewAI. kernel vs crew framework.
The two tools sit at different layers of the agent stack. The short version: Namzu is an agent kernel — runtime, scheduling, sandboxing. CrewAI is a multi-agent crew framework. They compose more often than they compete.
Namzu
Namzu is an open-source TypeScript agent kernel and SDK. It is unopinionated about how agents collaborate, what roles they take, or how they delegate. Those decisions live one layer up.
CrewAI
CrewAI is a Python framework for orchestrating teams of agents with explicit roles, goals, and delegation patterns. It models multi-agent work as a "crew" with sequential or hierarchical processes, and ships its own tooling around tasks, tools, and memory.
Choose Namzu if…
- →You are building in TypeScript, not Python.
- →You need OS-level isolation between concurrent agents — not just logical delegation.
- →You want to design your own coordination patterns rather than adopt a fixed crew/role abstraction.
- →You need checkpoint/resume across an agent run that may span hours.
Choose CrewAI if…
- →You want a ready-made vocabulary for roles, tasks, and crews.
- →Python is the team language and you value the broader Python AI ecosystem.
- →A simple sequential or hierarchical crew model fits your workload.
CrewAI describes who does what; Namzu describes how each one runs. They sit at different layers and do not overlap directly. A team that wants the CrewAI mental model in TypeScript has tended to write a thin coordinator on top of a runtime — that runtime is what Namzu provides.
Try the kernel underneath your stack.
Namzu installs from npm and runs on Node.js today. Pair it with whichever agent layer you already use.