Skip to main content

1.3 Feature entries

Current navigation groups common tools under Agent workbench, Device & development, and AI capabilities. The left sidebar can also show projects and conversations.

EntryPurposeManual
Agent WorkspaceSubmit tasks, choose a workflow, and resume project conversations3.1 Workbench
File managerLocal or board workspace files, transfers, and editing3.4 Files
TerminalLocal terminals, SSH terminals, and supported serial debugging3.3 Terminal
FlasherSelect board, image, and write target3.7 Flashing
Block studioChoose a robot, assemble blocks, simulate, and run on a device3.17 Block studio
Remote desktopUse the target device's graphical desktop3.6 Remote desktop
Code editorOpen a remote code workspace3.5 Code editor
Board AgentDeploy, configure, and use OpenClaw3.10 OpenClaw
Plugin CenterManage skills, MCP, and other extensions3.11 Plugin Center and skills
Local modelsInstall Ollama, download a model, and set the Agent main model3.12 Local models
SettingsAccount, devices, appearance, model/runtime, MCP, storage, and updates3.13 Configuration

Entries depend on deployment capabilities. Public web deployments do not directly provide disk writing, Ollama, or system integration on your computer; use a supported bridge flow or the desktop client. The skill store requires its service to be enabled and is normally reached from Plugin Center.

Common choices​

  • Ask a question: open the workbench; a board is optional.
  • Edit code on your computer: create a local project and bind its folder.
  • Run a board application: choose the device and board folder, then verify SSH is online.
  • Explore steps first: choose Plan. For approvable requirements and acceptance criteria, choose Spec.
  • Use DSH: check availability in settings, set it as the default for new tasks, then create a task.

The CLI is for scripts and automation; consult command help for supported capabilities. It does not provide graphical flashing wizards, a block canvas, or a remote desktop.