Build
OpenClaw AI Engine
Upgrade your workspace to next-generation infrastructure with isolated agents, multi-model support, and advanced capabilities.
What is OpenClaw?
OpenClaw is Pickaxe's next-generation AI engine that replaces the standard backend infrastructure powering your workspace. When enabled, every Pickaxe in your workspace runs as an independent OpenClaw instance — giving each user their own isolated, fully-featured AI agent operating on its own computer.
Think of it as upgrading from a shared environment to dedicated, sandboxed agents with significantly more capabilities out of the box.
How it works
When you enable OpenClaw at the workspace level, every Pickaxe switches from the default Pickaxe backend to OpenClaw routing. Each end user who interacts with one of your Pickaxes gets a unique OpenClaw instance running independently. This means:
- Isolated execution — every user session operates in its own sandboxed environment
- Multi-model functionality — OpenClaw agents can leverage multiple models within a single interaction
- Advanced actions — agents have access to a broader set of built-in capabilities beyond standard Pickaxe actions
- Deeper memory — improved context retention across conversations
Key features
Built-in capabilities
OpenClaw agents come with significantly more functionality enabled by default compared to standard Pickaxes:
- Web browsing — agents can search and browse the web in real time
- PDF creation — generate documents on the fly during a conversation
- Code execution — run code directly from the agent's sandboxed computer
- Computer access — request a link to view and control the agent's computer, allowing you or your users to authenticate accounts, verify actions, or interact with the environment directly
Rewritable agent logic
With OpenClaw, users have the ability to rewrite the brains of the agent. Pickaxe adds additional safeguards and guardrails on top to help prevent users from accessing or modifying things they shouldn't. This gives you flexibility to deeply customize agent behavior while maintaining security boundaries.
Enabling OpenClaw
To enable OpenClaw for your workspace:
- Navigate to your Workspace Settings
- Find the OpenClaw AI Engine toggle
- Turn it on — all Pickaxes in the workspace will now run on OpenClaw infrastructure
The switch applies workspace-wide, so every Pickaxe under that workspace will use OpenClaw routing once enabled.
Configuration
You have full control over your OpenClaw setup:
- Upload a configuration — bring your own OpenClaw configuration file and apply it to your workspace
- Download configurations — export the configuration from any of your Pickaxes to reuse, share, or modify externally
This makes it easy to version, replicate, or migrate your agent setups across workspaces.
Beta notice
OpenClaw is currently in beta. While it's fully functional, there is a lot more functionality still to come. During the beta period:
- Test in a deployment — we recommend testing your Pickaxes through a deployment (direct link, embed, portal, etc.) rather than through the builder to get the most accurate experience
- Expect updates — new capabilities and refinements will be rolled out regularly
- Share feedback — your input during the beta helps shape the future of OpenClaw
Getting started
Ready to try it? Head to your workspace settings, flip the toggle, and deploy a Pickaxe to see OpenClaw in action. If you run into anything unexpected, remember to test in a deployment environment for the most reliable results.