Skip to main content

3.2 AI Dock

The AI Dock is Moss’s input area. Open it from Core → Workbench, then follow confirm device → choose mode → enter your question → review results.

First-time flow

StepWhat to do
1Open Core → Workbench and find the input area at the bottom
2Check the device chip in the input area and confirm it points to the right device
3Choose Plan when impact is unclear; choose Execute when you want direct action
4Use Fast for simple questions; use Think for troubleshooting and code work
5After sending, check the chat, terminal, files, or diagnostics for outcomes

Common controls

ControlOptionsRole
Device chipCurrent device / not boundTells Moss whether the task can reach a device directly
Work modeExecute / PlanExecute runs the task; Plan lists steps, risks, and confirmation points first
Reply modeFast / ThinkFast suits short Q&A; Think suits complex troubleshooting and multi-step tasks
AttachmentsFiles / images / screenshotsGive Moss logs, configs, screenshots, or code snippets for context

Relationship to the workspace

Moss sits in the main column on the left; workspace panels are on the right. You can let Moss analyze while you open Terminal, Files, Changes, Diagnostics, or History. Moss uses the current page, device, project directory, and references you selected.

Offline and model error banners

The input area can show two important banners:

  • Device offline: Moss can still plan and answer knowledge questions; device execution waits for reconnect or confirmation.
  • Model configuration error: When Fast or Think points to local Ollama but the service is down or the model is missing, the banner links to Local LLM or AI model settings.

History and slash commands

The AI Dock keeps local chat history—you can start a new session or resume an older one. The input supports a few slash commands for clearing context, switching models, or viewing help; natural language is fine whenever you are unsure.

Next