⚡ What is Crawbl?
Crawbl is the control plane behind agent products. It gives users one stable API, gives each user a private runtime, and keeps the whole platform cheap enough to operate at scale.
This page is the fast mental model. In one short pass, you will see the main parts, what each part is responsible for, and where to go next in the docs.
🧱 The building blocks
Read the platform as a chain of responsibility: users enter through clients, the orchestrator makes decisions, ZeroClaw does the agent work, and outside systems provide models and tools when needed.
📱 Clients
The mobile app and future API consumers all enter through the same front door.
🧠 Orchestrator
It decides where requests go, what state changes, and which tools are allowed.
🤖 ZeroClaw
Each user runtime keeps its own workspace, sessions, and tool execution context.
🌐 Services
LLMs, Gmail, Slack, search, and other systems stay outside the control plane.
🚪 One front door
Users do not need to understand pods, runtimes, or agent internals. They talk to one API, and Crawbl hides the moving parts behind that entry point.
🧠 One control plane
The orchestrator is the brain of the platform. It handles auth, routing, persistence, integrations, MCP tools, and the rules for how messages move through the system.
🤖 One runtime per user
Each user gets an isolated ZeroClaw runtime. That separation protects state, keeps noisy users from affecting others, and lets runtime capacity scale on its own.
🧭 Where to go next
This page is the short tour. The next three pages turn that overview into real working context.
Tour complete
You have the shape of the product now. Use the next three pages to go deeper into architecture, trust boundaries, and deployment.